360
Relativistic Quantum Chemistry Applied to Actinides by Samuel Omonigho Odoh A Thesis submitted to the Faculty of Graduate Studies of The University of Manitoba in partial fulfillment of the requirements of the degree of DOCTOR OF PHILOSOPHY Department of Chemistry University of Manitoba Winnipeg, Canada Copyright © 2012 by Samuel O. Odoh

Relativistic Quantum Chemistry Applied to Actinides

  • Upload
    others

  • View
    3

  • Download
    0

Embed Size (px)

Citation preview

Page 1: Relativistic Quantum Chemistry Applied to Actinides

Relativistic Quantum Chemistry Applied to Actinides

by

Samuel Omonigho Odoh

A Thesis submitted to the Faculty of Graduate Studies of

The University of Manitoba

in partial fulfillment of the requirements of the degree of

DOCTOR OF PHILOSOPHY

Department of Chemistry

University of Manitoba

Winnipeg, Canada

Copyright © 2012 by Samuel O. Odoh

Page 2: Relativistic Quantum Chemistry Applied to Actinides

i

Table of Contents

Table of Contents ………………………………………………………………………………...i

List of Figures ……………………………………………………………………………….......iv

List of Tables ……………………………………………………………...…………………….xi

List of Abbreviations …………………………………………………………………………..xvii

Abstract …………………………………………………………………….…………………..xix

Acknowledgement ……………………………………………………………………………..xxi

Chapter 1- Introduction ……………………………………………..………….………………...1

1.1. The Actinide Elements and their Uses ……………………………….…….………………..1

1.2. Chemical Properties of Actinides ……………...………………..…………………………..4

1.3. Theoretical Studies of Actinide Complexes ….…………………..…………………….…...9

1.3.1. The Schrödinger Equation ….…………………..………………..…………….………….9

1.3.2. The Variational Principle and Electronic Basis Sets….………….……………………….10

1.3.3. The Hartree-Fock Method and Post Hartree-Fock Approaches …..…………...…………13

1.3.4. Density Functional Theory ….…………………………………….………….…………..17

1.3.5. Relativistic Effects ….……………………………….……………….…………………..21

1.3.6. Solvation Effects ….………………………………….…………….…………………….25

Page 3: Relativistic Quantum Chemistry Applied to Actinides

ii

1.4. Organization of this Thesis ….……………………………………….……..……………...26

1.5. References ....…….…………………………...…….……………………....……………....29

Chapter 2- Performance of Relativistic Effective Core Potentials for DFT Calculations on

Actinide Compounds ….……………………………………….……………….………………36

Chapter 3- Theoretical Study of the Structural Properties of Plutonium IV and VI Complexes

….……………………………………….……………………………………….……………...64

Chapter 4- Theoretical Study of the Structural and Electronic Properties of Plutonyl Hydroxides

….……………………………………….………………………………………………………98

Chapter 5- DFT Study of Uranyl Peroxo Complexes with H2O, F-, OH

-, CO3

2- and NO3

- ….151

Chapter 6- Theoretical Study of the Reduction of Uranium(VI) Aquo Complexes on Titania

Particles and by Alcohols …………………….……….………..……………………………...195

Chapter 7- QM and QM/MM Study of Uranyl Fluorides in the Gas Phase, Aqueous Phase and in

the Hydrophobic Cavities of Tetrabrachion …………………………………………………...231

Chapter 8- Novel and Stable U(V)/U(V) Binuclear Complexes Formed by Oxo-functionalization

of Axial Oxo Atoms …………………………………………………………………………..269

Chapter 9- Theoretical Study of a Gas-Phase Binuclear Uranyl Hydroxo Complex,

[(UO2)2(OH)5]- …………………………….………………………………………………….295

Chapter 10- Summary and Future Studies of Actinide Complexes …………………………..325

10.1. Summary ………………………………………………………………………………...325

Page 4: Relativistic Quantum Chemistry Applied to Actinides

iii

10.2. Future Studies of Actinide Complexes Specific to this Thesis …………..……………...328

10.3. General Directions for Computational Studies of Actinide Complexes ………………...330

References ……………….……………….……………….……………….…………………..331

List of Publications ……………………………………………………………………………336

Page 5: Relativistic Quantum Chemistry Applied to Actinides

iv

List of Figures

Figure 1.1: The actinides in the periodic table of elements ……………………………..……….2

Figure 1.2: The aquo complexes of U(VI) and U(IV) …………………………………..………8

Figure 1.3: Schematic description of the chapters in this thesis ………………………….…….28

Figure 2.1: Absolute Deviations of the Vibrational Wavenumbers, cm-1

for NUN, NUO+ and

CUO computed using Stuttgart RECPs with the aug-cc-pVTZ ligand basis set and the AE four-

component method with the L2 basis set ………………………………………………….…...44

Figure 2.2: Distorted Planar T-Shaped Structure of UO3(g) ……………………………….…..46

Figure 2.3: Absolute Deviations of the Vibrational Wavenumbers, cm-1

, for UO3 and UF6

computed using Stuttgart RECPs with the aug-cc-pVTZ ligand basis set and the AE four-

component method with the L2 basis set ………………………………………………….…..50

Figure 2.4: Absolute Deviations of the Enthalpy Changes, kJ/mol for Reactions 1-5 computed

using Stuttgart RECPs with the aug-cc-pVTZ ligand basis set and the AE four-component

method …………………………………………………………………………………….…...55

Figure 3.1: (Left) The An=Oyl bond lengths (Å) in the bare actinyl moieties, AnO22+

as well as

the An=Oyl and average An-OH2 bond lengths (Å) in the pentaaquo complexes, [AnO2(H2O)5]2+

.

(Right) The actinyl symmetric and asymmetric stretching vibrational frequencies (cm-1

) in the

bare actinyl moieties, AnO22+

and the pentaaquo complexes, [AnO2(H2O)5]2+

. ………….…...74

Figure 3.2: (Left) The An=Oyl bond lengths (Å) in the bare actinyl moieties, AnO21+

as well as

the An=Oyl bond lengths and average An-OH2 bond lengths (Å) in the pentaaquo complexes,

Page 6: Relativistic Quantum Chemistry Applied to Actinides

v

[AnO2(H2O)5]1+

. (Right) The actinyl symmetric and asymmetric stretching vibrational

frequencies (cm-1

) in the bare actinyl moieties, AnO21+

and the pentaaquo complexes,

[AnO2(H2O)5]1+

. These values were obtained with the obtained using the PBE functional and

PCM solvation model ……………………………………………………………………….….75

Figure 3.3: A. The cis (top) and trans (bottom) isomers of the PuO2Cl2(H2O)2 complex. B. The

cis (top) and trans (bottom) isomers of the PuO2Cl2(H2O)3 complex …………………………80

Figure 3.4: A) The decrease in the An=Oyl and An-Cl bond lengths (Å) on addition of 2Cs+ to the

[AnO2Cl4]2-

complexes. B) The increase in the actinyl stretching vibrational frequencies (cm-1

)

on addition of 2Cs+ to the [AnO2Cl4]

2- complexes …………………………………………….81

Figure 3.5: The optimized structure of plutonyl diaquo-dinitrate, PuO2(NO3)2(H2O)2 obtained

using the B3LYP functional ……………………………………………………………………85

Figure 3.6: The optimized structures of several plutonium (IV) complexes: A) Pu(H2O)84+

, B)

Pu(NO3)22+

, C) Pu(NO3)4(H2O)3 and D) Pu(NO3)62-

calculated with the B3LYP functional in the

aqueous phase …………………………………………………………………………………..89

Figure 4.1: Optimized structures of the aquo-hydroxo [PuO2(H2O)5-n(OH)n]2-n

complexes

obtained at the BP86/B2 level with the COSMO solvation model ……………..……….……105

Figure 4.2: Optimized structures of the aquo-hydroxo [PuO2(H2O)4-n(OH)n]2-n

complexes

obtained at the BP86/B2 level with the COSMO solvation model …………………………...106

Figure 4.3: Optimized structures of the bis-plutonyl aquo tetrahydroxo and dihydroxo complexes

obtained at the B3LYP/B3 level in the gas phase …………...………………………………..114

Page 7: Relativistic Quantum Chemistry Applied to Actinides

vi

Figure 4.4: Structures of [(PuO2)3(H2O)6(O)(OH)3]+ obtained at the BP86/B2/COSMO level. The

C3V structure is shown on the left while the structure with a bridging aquo group is shown on the

right ……………………………………………………………………………………………115

Figure 4.5: Scheme depicting the formation and decomposition of the μ3-oxo hexagonal

trimetallic core of the trimer aquo-hydroxo complexes ………….…………………………....127

Figure 4.6: The molecular orbitals of [PuO2]2+

...…………….……..……………...………....129

Figure 4.7: The energy levels of the [AnO2]2+

optimized at the BP86/B2 level with the COSMO

model ………………………………………………………………………………………….130

Figure 4.8: Selected occupied alpha spin MOs of [PuO2(OH)4]2-

……………………..……..132

Figure 4.9: The energy levels of the [AnO2(OH)4]2+

optimized at the BP86/B2 level with the

COSMO model ……………………………………………….……………………………….134

Figure 4.10: The 39th to 62nd MOs of [(PuO2)2(OH)2]2+

of alpha spin ……………..……….136

Figure 4.11: Selected alpha spin MOs of the trimer complex, [(PuO2)3(O)(OH)3]+ …..……...139

Figure 4.12: Abbreviated energy level diagram of the [(AnO2)3(O)(OH)3]+

complexes ……...142

Figure 5.1: The structures of UO22+

and its peroxo derivatives optimized at the B3LYP/B1 level

in aqueous solution ……………………………………….………………...…………………156

Figure 5.2: The molecular orbitals of UO2(O2). The geometry of this complex was optimized

with the PCM approach and the B3LYP functional ………………………………………….159

Page 8: Relativistic Quantum Chemistry Applied to Actinides

vii

Figure 5.3: Simulated IR spectra of UO22+

and its peroxo derivatives obtained at the at the

B3LYP/B1 level in aqueous solution ………………………………………………………….160

Figure 5.4: MO-26 and MO-27 of the C2v structure of UO2(O2)22-

………………………….163

Figure 5.5: The structures of UO2(H2O)52+

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution ………………………………………………………………………166

Figure 5.6: The structures of UO2F42-

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution ………………………………………………………………………170

Figure 5.7: The structures of UO2(OH)42-

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution ………….……………………………………...……………………173

Figure 5.8: MO-30 of cis and trans- UO2(O2)2(OH)24-

………….………..…………………..176

Figure 5.9: The structures of UO2(CO3)34-

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution ……………………………………………………………………....178

Figure 6.1: Possibilities for charge transfer to U(VI) complexes adsorbed on TiO2 crystals or

nanoparticles …………………………………………………………………………………..199

Figure 6.2: Structure of UO22+

adsorbed at the rutile (110) surface at the bridging oxygen atoms

…………………………...............…………………………………………………………….204

Figure 6.3: Total and partial electronic density of states (DOS) obtained for a clean rutile (110)

slab while using with the PBE functional (left) and the PBE+U approach (right, U=4.2 eV)

…………………………………………………………………………………………………206

Page 9: Relativistic Quantum Chemistry Applied to Actinides

viii

Figure 6.4: Total and partial electronic density of states (DOS) obtained for hydroxylated rutile

(110) slabs with the PBE+U approach (U=4.2 eV for Ti 3d and U=4.0 eV for U 5f)

……………………………………….………………………………………………………...207

Figure 6.5: Total and partial electronic density of states (DOS) obtained for rutile (110) slabs

with an adsorbed UO22+

group obtained with the PBE+U approach (U=4.2 eV for Ti-3d and

U=4.0 eV for U-5f)….………………………………………………………………………...210

Figure 6.6: Total and partial electronic density of states (DOS) obtained for stoichiometric (left)

and surface hydroxyl defected (right) rutile (110) slabs with adsorbed [UO2(H2O)3]2+

obtained

with the PBE+U approach (U=4.2 eV for Ti-3d and U=4.0 eV for U-5f) ………………….211

Figure 6.7: Electronic energy levels and frontier molecular orbitals obtained at the BP86/TZP

level for the a Ti38076-[UO2(H2O)3]2+

surface-adsorbate complex …………………………...212

Figure 6.8: Calculated spin distributions of a rutile (110) slab with a surface-adsorbed

[UO2(H2O)3]2+

moiety and a surface hydroxyl defect. These calculations were carried out with

the PBE+U approach …………………………………………………………………………215

Figure 6.9: Electronic energy levels and MOs of a non-stoichiometric cluster (with a surface

hydroxyl)-adsorbate complex …….......………………………………………………………216

Figure 6.10: The quenching of the lowest triplet excited state of [UO2(H2O)5]2+

by water and

organic alcohols ………………………….…………………………………………………...221

Figure 7.1: Top: the right handed coiled coil protein of Tetrabrachion. Bottom: two monomer

chains of the RHCC tetramer ………………………………………………………………...233

Page 10: Relativistic Quantum Chemistry Applied to Actinides

ix

Figure 7.2: Aqueous phase structures of the [UO2Fn(H2O)5-n]2-n

complexes optimized at the

BP86/TZP/ZORA/COSMO level ……………………………………………………………..240

Figure 7.3: Left: Variation of the Mulliken charges on the uranium atom (black) and uranyl

moiety (red) and the fluoro-2p contribution to the HOMO-1 orbital (blue) with increasing

number of fluoride ligands. Right: Variation of the U=O bond lengths (black) and bond orders

(red) with increasing number of fluoride ligands ……………………………………………..246

Figure 7.4: Selected frontier molecular orbitals of [UO2Fn(H2O)5-n]2-n

complexes. The member

structures were optimized at the ADF/ZORA/BP86/TZP/COSMO ……………….………....248

Figure 7.5: The atoms of the isoleucine residues (represented as green crosses) at the n-terminal

end of cavity two ....…………………………………………………………………………..255

Figure 8.1: Two common motifs for cation-cation interactions between uranyl groups ..........272

Figure 8.2: The recently synthesized U2O4 Pacman complex, 1a …………............................273

Figure 8.3: Optimized structure of the yttrium dimer complex obtained at the PBE/L1 level. The

O, U and Y atoms are in red, dark green and light green colours respectively ………………274

Figure 8.4: The lithium-chloride monomer salt of the yttrium dimer complex, 2. The dioxo-

uranium core with Li and Y oxo-functionalization are shown on the right ….………………275

Figure 8.5: Molecular orbitals of primary σ- and π- character in the unrestricted singlet state of

1a ….………………………………………………………………………………………….280

Figure 8.6: The (a) HOMO-172 (bonding with respect to the two U atoms) with an energy of -

1.094 a.u. and (b) its antibonding counterpart, both containing contributions from U-5f and O-2s

orbitals ………………………………………………………………………………………..281

Page 11: Relativistic Quantum Chemistry Applied to Actinides

x

Figure 9.1: The three low energy structures of (UO2)2(OH)5- obtained at the B3LYP/TZVP level

………………………………………………………………………………………………...301

Figure 9.2: Calculated IR spectra for the μ2-dihydroxo (black), μ-hydroxo-CCI (blue) and μ-

hydroxo-di-CCI (red) structures of (UO2)2(OH)5- obtained using the B3LYP functional

………………………………………………………………………………………………...308

Figure 9.3: Other possible structure of (UO2)2(OH)5-. Their energies at the B3LYP (and MP2

//B3LYP) levels are given relative to that of the μ-hydroxo-di-CCI structure

………………………………………………………………………………………………...313

Figure 9.4: Low energy structures of the bis-uranyl hydroxo complexes, (UO2)2(OH)n4-n

………………………………………………………………………………………………...316

Page 12: Relativistic Quantum Chemistry Applied to Actinides

xi

List of Tables

Table 1.1: Electronic ground state configurations of the actinide elements ……………………..5

Table 1.2: The various oxidation states of the actinide elements ………………………………..6

Table 2.1: Vibrational Frequencies, cm-1

, of NUN computed using Stuttgart RECPs with the aug-

cc-pVTZ ligand basis set and the AE four-component method with the L2 basis set …………42

Table 2.2: Vibrational Frequencies, cm-1

, of NUO+ computed using Stuttgart RECPs with the

aug-cc-pVTZ ligand basis set and the AE four-component method with the L2 basis set….….43

Table 2.3: Vibrational Frequencies, cm-1

, of CUO computed using Stuttgart RECPs with the aug-

cc-pVTZ ligand basis set and the AE four-component method with the L2 basis set ………...45

Table 2.4: Calculated Bond Lengths (in Ǻ) and Bond Angles (in degrees) of UO3 (g) computed

using Stuttgart RECPs with the aug-cc-pVTZ ligand basis set and the AE four-component

method with the L2 basis set …………………………………………………………………...48

Table 2.5: Calculated Bond Lengths (in Ǻ) and Bond Angles (in degrees) of UO3 (g) computed

using Stuttgart RECPs with the aug-cc-pVTZ ligand basis set and the AE four-component

method with the L2 basis set ………………………………………………..…………….……49

Table 2.6: Vibrational Frequencies, cm-1

, of UF6 computed using Stuttgart RECPs with the aug-

cc-pVTZ ligand basis set and the AE four-component method with the L2 basis set …………51

Table 2.7: Calculated enthalpy changes*, kJ/mol for Reactions 1-5 computed using Stuttgart

RECPs and the AE four-component method …………………………………………………..53

Page 13: Relativistic Quantum Chemistry Applied to Actinides

xii

Table 3.1: The calculated bond lengths (Å) and plutonyl vibrational frequencies (cm-1

) of the

bare PuO22/1+

and actinyl aquo PuO2(H2O)52/1+

systems obtained using DFT in the gaseous and

aqueous phases ………………………………………………………………………………….71

Table 3.2: The calculated structural properties (bond lengths in Å and vibrational frequencies in

cm-1

) of the chloride complexes of the plutonyl (VI) cation obtained using the B3LYP functional.

The values obtained with the PBE functional are given in parenthesis ...………………………77

Table 3.3. The calculated structural properties (bond lengths in Å and vibrational frequencies in

cm-1

) of the nitrate complexes of the plutonyl (VI) cation obtained using the B3LYP functional

in the gaseous and aqueous phases. The values obtained with the PBE functional are given in

parenthesis ………………………………..…………………………….……….………………84

Table 3.4: The calculated bond lengths (Å) of the AnO2(NO3)2(H2O)2, AnO2(NO3)2(TBP)2a,

AnO2Cl2(H2O)2 and AnO2Cl2(TPPO)2a complexes obtained using the B3LYP functional with

RECPs in the gaseous phase ……………………………………………………………………86

Table 3.5: The calculated Pu-OH2 and Pu-Owater/nitrate bond lengths of the aquo, nitrate and aquo-

nitrate complexes of the plutonium (IV) cation obtained using DFT in the gaseous and aqueous

phases …………………………………………………………...………………………………88

Table 4.1: Calculated structural properties of [PuO2(H2O)5]2+

and [PuO2(H2O)4(OH)]+. The bond

distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

…………...……………………………..………………..106

Page 14: Relativistic Quantum Chemistry Applied to Actinides

xiii

Table 4.2: Calculated structural properties of [PuO2(H2O)3(OH)2] and [PuO2(H2O)2(OH)3]-. The

bond distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

…………………………………………………………...108

Table 4.3: Calculated structural properties of [PuO2(H2O)2(OH)2] and [PuO2(H2O)(OH)3]-. The

bond distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

…………………………………………………………...109

Table 4.4: Calculated structural properties of [PuO2(OH)4]2-

and [PuO2(OH)5]3-

. The bond

distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

…………………………………………………………...110

Table 4.5: Calculated Mayer bond orders of the plutonyl aquo-hydroxo complexes obtained at

the PBE/B1 level in the gaseous phase …..……….…………………………………………...111

Table 4.6: Calculated structural properties of the plutonyl dimer complexes. The bond lengths

are in Å and the calculated IR intensities (km/mol) are given in parenthesis …………………113

Table 4.7: Calculated structural properties of the μ3-oxo motifs of the trimeric complex,

[(PuO2)3(H2O)6(O)(OH)3]+. The bond lengths are given in Å, bond angles in degrees while

vibrational frequencies are presented in cm-1

…………………..……………………………..115

Table 4.8: Calculated uranyl stretching vibrational frequencies (cm-1

) of uranium aquo hydroxo

complexes obtained at the PBE/B1 level ……………………………………………………...118

Table 4.9: Calculated reaction energies (kcal/mol) for the formation of the dimer and trimer

actinyl hydroxo complexes in the aqueous phase ……………………………………………..122

Page 15: Relativistic Quantum Chemistry Applied to Actinides

xiv

Table 4.10: Calculated reaction energies (kcal/mol) obtained actinyl hydrolysis reactions at the

BP86/B2 while using the COSMO solvation model ………………………………………….124

Table 4.11: Energies, atomic contributions and descriptions of the MOs of [(PuO2)2(OH)2]2+

at

the B3LYP/B3 level. The orbital energies are scaled such that the HOMO is at 0.00 eV …...137

Table 4.12: Energies (eV), individual atomic contributions and descriptions of the MOs of

[(PuO2)3(O)(OH)3]+ at the B3LYP/B3 level. The orbital energies are scaled such that the HOMO

is at 0.00 eV …………………………………………………………………………………...141

Table 5.1: Energies and characters of the MOs of the dioxouranium (VI) peroxides in aqueous

solution obtained at the B3LYP/B1 level. MO energies are given in eV …………………….158

Table 5.2: Calculated structural properties and vibrational frequencies of UO22+

and its peroxo

derivatives obtained at the B3LYP/B1 level in aqueous solution ………………………….....161

Table 5.3: Calculated structural properties and vibrational frequencies (cm-1

) of UO2(H2O)52+

and

its peroxo derivatives obtained at the B3LYP/B1 level in the gas-phase and in aqueous solution

…………………………………………………………………………………………………165

Table 5.4: Calculated structural properties and vibrational frequencies of UO2F42-

and its peroxo

derivatives obtained at the B3LYP/B1 level in aqueous solution …………………………….168

Table 5.5: Calculated structural properties and vibrational frequencies of UO2(OH)42-

and its

peroxo derivatives obtained at the B3LYP/B1 level in aqueous solution …………………….172

Table 5.6: Calculated structural properties and vibrational frequencies (cm-1

) of UO2(CO3)34-

and

its peroxo derivatives obtained at the B3LYP/B1 level in the gas-phase and in aqueous solution

…………………………………………………………………………………………………180

Page 16: Relativistic Quantum Chemistry Applied to Actinides

xv

Table 5.7: Calculated structural properties and vibrational frequencies (cm-1

) of UO2(NO3)3- and

its peroxo derivatives obtained at the B3LYP/B1 level in the gas-phase and in aqueous solution

…………………………………………………………………………………………………181

Table 5.8: The calculated Mayer bond orders in various uranyl complexes and their peroxo

derivatives obtained at the B3LYP/B2 level using structures optimized at the B3LYP/B1 level.

……………………………………………………………………………………………..…..184

Table 6.1: Calculated structural features and properties of U(VI) complexes obtained with the

PBE functional in the gas phase ……………………………………………………………....203

Table 6.2: Calculated reaction energies (kcal/mol) obtained for R6 in aqueous solution obtained

at the PBE/TZP and B3LYP/TZP levels in addition to the calculated structural and electronic

properties of the triplet state uranyl-quencher complexes ……………………………………219

Table 7.1: Calculated and experimental structural parameters of the [UO2Fn(H2O)5-n]2-n

complexes …………………………………………………...…………………………...........241

Table 7.2: Aqueous phase calculated ligand binding energies in kcal/mol, Mayer bond orders for

the U=O Bond and atomic Mulliken charges on uranium atoms in the UO2Fn(H2O)5-n]2-n

complexes obtained at the ADF/ZORA/TZP/BP86/COSMO level ………………………….247

Table 7.3: Calculated relative energies (kcal/mol), frontier gaps (eV) and structural features of

the two orientations of the [UO2F5]3-

complex in the hydrophobic cavities of tetrabrachion...256

Table 7.4: Computed structural parameters cisplatin in the gaseous and aqueous phases and of

two randomly selected orientations of cisplatin embedded in the cavity two (largest cavity) of the

RHCC protein obtained using RECPs ………………………………………………………..260

Page 17: Relativistic Quantum Chemistry Applied to Actinides

xvi

Table 8.1: Calculated and experimental structural parameters of 1a in the ferromagnetic triplet

electronic state and antiferromagnetic broken-symmetry state (in parentheses) ……………...279

Table 8.2: Calculated reaction free energies (ΔG298, kcal/mol) required to transaminate the

unoccupied amine site of several Pacman complexes with a cis-uranyl group ………..……...284

Table 8.3: Calculated atomic charges on the uranium and oxo- atoms of the uranyl Pacman

complex, UO2(Py)(H2L), and its pentavalent and reductively oxo-functionalized derivatives

obtained at the PBE/AE/4-component level (and at the B3LYP/RECP level) ……………….284

Table 8.4: The calculated and experimental structural parameters (bond lengths in Å and angles

in degrees) for the monomer complexes ………………............................................................286

Table 8.5: The calculated and experimental structural parameters (bond lengths in Å and angles

in degrees) for the dimer complexes ………………...………………………………………...288

Table 9.1: Calculated relative energies (kcal/mol) of the three low energy structures of

(UO2)2(OH)5- ………………………………………………………………………………….302

Table 9.2. Calculated bond lengths (Å) and bond orders of the low energy structures of

(UO2)2(OH)5- obtained at the B3LYP/TZVP level …………………………………………...303

Table 9.3. Calculated IR vibrational frequencies (cm-1

) of the low energy structures of

(UO2)2(OH)5- obtained at the B3LYP/TZVP level …………………………………………...307

Table 9.4. Calculated bond lengths (Å), bond orders and vibrational frequencies (cm-1

) of a few

gas-phase mono-nuclear uranium complexes obtained at the B3LYP/TZVP level …………..310

Table 9.5. Relative energies (kcal/mol) of the low energy structures of (UO2)2(OH)n4-n

, (n=2, 3, 4

and 6) obtained at the DFT and ab initio levels ……………………………………………….315

Page 18: Relativistic Quantum Chemistry Applied to Actinides

xvii

List of Abbreviations

AE All-Electron

BO Born-Oppenheimer

CASSCF Complete Active Space Self Consistent Field

CASPT2 CASSCF with 2nd

-Order Perturbative Treatment of Dynamical Correlation

CCI Cation-Cation Interactions

CCSD Coupled Cluster Singles and Doubles

CCSD(T) Coupled Cluster Singles and Doubles with Perturbative Triples

CCSDTQ Coupled Cluster Singles, Doubles, Triples and Quadruples

CI Configuration Interaction

DKH Douglas Kroll Hess

DFT Density Functional Theory

ECP Effective Core Potentials

FW Foldy-Wouthuysen

GGA Generalized Gradient Approximations

GTOs Gaussian Type Orbitals

HF Hartree-Fock

HK Hohenberg-Kohn

HLNW High Level Nuclear Waste

HOMO Highest Occupied Molecular Orbitals

IR Infra-Red

KS Kohn-Sham

LCAO Linear Combination of Atomic Orbitals

LC-ECP Large Core Effective Core Potentials

Page 19: Relativistic Quantum Chemistry Applied to Actinides

xviii

LDA Local Density Approximation

LUMO Lowest Occupied Molecular Orbitals

MO Molecular Orbital

MPn Møller-Plesset Perturbation Theory of nth Order

NBO Natural Bond Orbitals

NMR Nuclear Magnetic Resonance

NOCV Natural Orbitals for Chemical Valence

PBE Perdew Burke and Enzerhof GGA Functional

PCM Polarizable Continuum Model

Post HF Post HF approaches

PUREX Plutonium URanium EXtraction

RECP Relativistic Effective Core Potentials

SCF Self Consistent Field

SC-ECP Small-Core Effective Core Potentials

STOs Slater Type Orbitals

UA0 United Atom Topological Model

UEG Uniform Electron Gas

WIPP Waste Isolation Pilot Plant

ZORA Zeroth Order Regular Approximation

Page 20: Relativistic Quantum Chemistry Applied to Actinides

xix

Abstract

Of the many available computational approaches, density functional theory is the most

widely used in studying actinide complexes. This is generally because it incorporates electron

correlation effects and is computationally inexpensive for modestly sized compounds.

The first chapter of this thesis is an introductory chapter in which some basic concepts of

electronic structure theory are discussed. The rest of this thesis is a compilation of several studies

of the structural and electronic properties of a range of actinide compounds using predominantly

density functional theory. The performances of the basis set/relativistic components as well as

the density functional component of theoretical calculations were examined in Chapters 2 and 3

respectively. In Chapters 4, 5, 6 and 7, the electronic structures and properties of actinide species

in the environment were explored. The speciation of actinyl aquo-hydroxo species at increasing

pH values were studied in Chapter 4. In Chapter 5, the structural and electronic properties of

uranyl peroxo complexes with other environmentally important ligands were studied. The

adsorption of uranyl complexes to geochemical surfaces was studied in Chapter 6. In addition,

the mechanistic pathways to the reduction of these complexes on surfaces and alcohols were

examined. In chapter 7, the complexes formed by the uranyl moiety with the aquo and fluoride

ligands were studied in gas and aqueous phases. The interactions of uranyl pentafluoride with a

protein were examined using a hybrid QM/MM approach. Overall these studies (Chapters 4, 5, 6

and 7) provided valuable insights into the speciation and reduction of actinide species in the

environment. In Chapter 8, the properties of novel pentavalent uranium complexes were studied

using density functional theory. These complexes have promising roles in the retardation of

uranium, via U(VI)-U(IV) reduction, in the environments of nuclear storage repositories. In

Page 21: Relativistic Quantum Chemistry Applied to Actinides

xx

Chapter 9, the existence of cation-cation interactions in an hexavalent bis-uranyl hydroxo

complex was examined using density functional theory and wavefunction methods. In Chapter

10, a summary of the works compiled in this thesis is presented. Future directions for work on

the chemistry of actinide complexes were also included in this chapter.

Page 22: Relativistic Quantum Chemistry Applied to Actinides

xxi

Acknowledgement

I would like to thank my supervisor, Dr Georg Schreckenbach, for the opportunity to

work under him towards this degree over the last few years. His patience, support,

encouragement, friendly approach and critical thinking have been invaluable along the way to

this degree. In the same vein, I would like to thank the members of my advisory committee, Dr

Mario Bieringer, Dr Peter Budzelaar and Dr Mostafa Fayek for their help, advice, criticism and

support during the work towards this degree.

I am indebted to all the present and former members of the Schreckenbach group at the

University of Manitoba. The postdoctoral and research fellows, Grigory Shamov, Qing-Jiang Pan

and Abu Asaduzamman were very helpful during the early stages of my doctoral program. Their

help with the various projects I worked on as well as on-going collaboration after their departure

from the group is highly appreciated.

I would like to thank all members of my family. Without the encouragement of my loving

wife, Sarah Odoh, my sister, Mary Odoh, and my parents (Godwin and Caroline Odoh), it would

not have been possible to complete this degree. I am eternally grateful to my brothers, Emmanuel

and Daniel Odoh for their support and understanding.

Page 23: Relativistic Quantum Chemistry Applied to Actinides

1

Chapter 1: Introduction

1.1 The actinides and their uses.

The actinides or actinoids are the elements with atomic numbers from 89 to 103 on the

periodic table of elements.1-2

They are named after the first member of this series, actinium. The

position of these elements on the periodic table is shown in Figure 1.1. The actinide series

correspond to filling of the 5f shell (mostly) and the actinides are therefore f-block elements. The

only exception to this is lawrencium which is a d-block element.3 The other members of the f-

block are the lanthanides. These are the elements from lanthanum to lutetium which possess

gradually filled 4f shells.

Of the actinides, only uranium and thorium are found in substantial quantities in nature.

Protactinium and actinium are also found in nature. All the other actinides are artificial elements

produced through various nuclear reactions of primordial uranium. In terms of abundance,

uranium and thorium exist at average concentrations of about 2-4 and 6 parts per million (ppm)

in the earth crust. Examples of thorium minerals are thorianite (ThO2, 88% Th), thorite (ThSiO4,

72% Th) and brabantite (CaTh(PO4)2, 50% Th). The most common uranium ore is uraninite

(UO2, 88% U). Others include Rutherfordine (UO2(CO3), 72% U) and schoepite

[(UO2)8O2(OH)12.12H2O, 73% U]. Np, Pu, Cm, Bk and Cf are sometimes formed by natural

transmutation of uranium ores and so are found in minute quantities in these minerals.4

The actinides have found their greatest applications in the production of energy via

controlled nuclear fission and in the production of nuclear weapons. These uses are underpinned

Page 24: Relativistic Quantum Chemistry Applied to Actinides

2

Figure 1.1: The actinides in the periodic table of elements.5

by the radioactive behavior of the actinide elements. Radioactivity is the ability of an unstable

nucleus to lose energy and decay to other nucleus/nuclei by emitting ionizing radiation. Although

all isotopes of the actinides are radioactive, the 235 isotope of uranium is the most commonly

used fissile material in nuclear reactors. Bombardment of this isotope with a neutron leads to its

fragmentation into smaller nuclei and emission of 2-3 other neutrons. As a result of the release of

more neutrons than were used in the initiation process, the fission of 235

U becomes self-

sustaining after a critical mass (about 52 kg) is attained. This chain-reaction of neutron-induced

fission of 235

U is controlled in nuclear reactors. The heat produced is used to generate electricity.

Currently, about 7% of total global energy consumption and 14% of global electricity

consumption is produced from nuclear reactors using some fissionable actinide isotope.6

Page 25: Relativistic Quantum Chemistry Applied to Actinides

3

There are however significant drawbacks to the use of the actinide elements in nuclear

reactors to generate energy.7 Firstly, significant portions of the nuclear fuel used in fission

reactors are not consumed and end up as waste, generally a mixture of U, Pu and other actinides.

Some constituents of the radioactive waste, especially fissile plutonium, can be separated and

reused in nuclear weapons and other reactors. The plutonium uranium extraction process,

PUREX, is one such method for extracting fissile actinides from spent nuclear fuel. It is however

the case that spent fuel and nuclear waste are highly hazardous and toxic to living things as well

as the environment. There is therefore a need to keep nuclear waste and spent fuel from

contaminating the environment. The half-life of a radionuclide (radioactive nucleus) is the period

of time it takes it to lose half of its radioactivity. The actinides with very long half-lifes found in

nuclear waste pose strong challenges to waste storage, disposal and management strategies. 232

U

has the shortest half-life (68.9 years) of the uranium isotopes while 238

U has the longest half-life

(4.5 billion years). 237

Np and 239

Pu are also found in nuclear waste and have half-lifes of about 2

million and 24,000 years respectively. The rather long periods of time needed for these

radioisotopes to lose their radioactivity, relative to the average human life span, means that it is

well nigh impossible to prevent eventual dispersal into the environment, waste management

approaches such as geologic disposal and transmutation notwithstanding. For this reason, it is

very important that we have a clear understanding of the dispersal and migration of actinide

elements, their speciation in the environment, their interaction with abiotic surfaces as well as

with biotic organisms and the mechanisms behind their toxicity to animal and plant life.8 There is

a need to understand the overall chemistry occurring at already contaminated sites (that are here

and now).

Page 26: Relativistic Quantum Chemistry Applied to Actinides

4

The actinides are also used in coloring glasses and ceramics (UO2)9, smoke detectors (Am

as alpha emission source)10

, and gas mantles (Th). 239

Pu is extensively used in the nuclear

weapons industry while 238

Pu is used in heart pacemakers and deep-sea diving suits as a source

of energy.11

It was used as heating source for the astronauts who participated in the Apollo space

missions. It is particularly suited for these roles as it emits relatively harmless alpha particles.

239Pu is now being used as a nuclear fuel in fast-breeder reactors. Depleted uranium is used in

making battle armors and projectiles. The ability of actinide elements and their compounds to

efficiently catalyze reactions is under continuing investigation.

1.2 Chemical properties of the actinides

There have been a large number of studies examining the chemical properties of the

actinide elements and their compounds. For the sake of brevity we here focus on their oxidation

states and electronic configurations. The electronic ground state configurations of the actinide

elements are given in Table 1.1. The 5f, 6d and 7s electrons are close in energy to each other as a

result of the relative destabilization of the 5f orbitals due to relativistic effects. The implication

of this is that the actinides can have a variety of oxidation states as any number of ionized

electrons can be removed from the energetically close valence energy levels. The various

oxidation states existing for each of the actinide elements are shown in Table 1.2. The presence

of multiple oxidation states for the actinides has significant ramifications to their speciation in

the environment as the stability and hydrolytic behavior of each actinide elements differ for

different oxidation states. This is particularly common for the light actinides (U, Np and Pu)

which in some cases exhibit more than one oxidation state in the same solution. As expected, the

stabilities of the different oxidation states depend on the electronic configuration (and resulting

stability) of the resulting ion. For example, the +2 oxidation state is known to be only transiently

Page 27: Relativistic Quantum Chemistry Applied to Actinides

5

Table 1.1: Electronic ground state configurations of the actinide elements. The noble gas core

structure of radon, [Rn], is used to depict the electronic configurations. The atomic number, Z, of

each element is also given.1

Element Z Configuration Element Z Configuration

Actinium 89 [Rn]6d17s

2 Berkelium 97 [Rn]5f

97s

2

Thorium 90 [Rn]6d27s

2 Californium 98 [Rn]5f

107s

2

Protactinium 91 [Rn]5f26d

17s

2 Einsteinium 99 [Rn]5f

1!7s

2

Uranium 92 [Rn]5f36d

17s

2 Fermium 100 [Rn]5f

127s

2

Neptunium 93 [Rn]5f46d

17s

2 Mendelevium 101 [Rn]5f

137s

2

Plutonium 94 [Rn]5f67s

2 Nobelium 102 [Rn]5f

147s

2

Americium 95 [Rn]5f77s

2 Lawrencium 103 [Rn]5f

146d

17s

2

Curium 96 [Rn]5f76d

17s

2

stable for all the actinide elements (with the exception of Nobelium, due to the presence of a

stable fully filled 5f sub-shell). Similarly the half-filled 5f7 electronic configuration implies that

the +4 oxidation state is most stable for Berkelium. The different oxidation states for these

elements result in a rich redox chemistry and consequently colorful chemistry in solution. For

example aqueous solutions of uranium in the +3, +4 and +6 oxidation states have visible colors

of brown-red, green and yellow respectively. The +4 and +6 oxidation states are most important

for uranium. They form the oxides, UO2 and UO3 respectively. The +5 oxidation state of

uranium is generally unstable as it disproportionates to the +4 and +6 oxidation states.12 The

different actinide oxidation states also generally have different solubility and stability in aqueous

Page 28: Relativistic Quantum Chemistry Applied to Actinides

6

Table 1.2: The various oxidation states of the actinide elements.

Actinium 3 Berkelium 3, 4

Thorium 3, 4 Californium 2, 3, 4

Protactinium 3, 4, 5 Einsteinium 2, 3

Uranium 3, 4 ,5 ,6 Fermium 2, 3

Neptunium 3, 4 ,5 ,6 ,7 Mendelevium 2, 3

Plutonium 3, 4 ,5 ,6 ,7 Nobelium 2, 3

Americium 3, 4 ,5 ,6 Lawrencium 3

Curium 3, 4, 5

environments.8 Uranium in the +6 state, U(VI), is generally more soluble in water than the +4

oxidation state, U(IV), which tends to form insoluble precipitates. To deter contamination of the

aquatic environment and the ecological zones of cities and countries by radioactive uranium

from spent fuel or mill tailings, it is preferable that U(VI) compounds are reduced to the +4

oxidation state. This is because the U(IV) compounds are significantly less soluble in water and

therefore migrate more slowly than U(VI) species and are more amenable to retardation and

deposition strategies.8 An interesting example of the role the differing oxidation states play in the

environmental chemistry of the actinide elements has been described by Runde.8 Plutonium is a

rather problematic element in the waste isolation pilot plant, WIPP, depository, in New Mexico,

USA. This is because the high chloride soil content of the depository preferentially stabilizes

Page 29: Relativistic Quantum Chemistry Applied to Actinides

7

Pu(VI) which is more soluble (and thus more environmentally mobile) than Pu(IV). In contrast,

the alkaline waters of the Yucca Mountain nuclear waste storage site dissolve Np about a

thousand times more easily than Pu. As such Np is the problematic element at Yucca Mountain,

Nevada, USA.8

The +3 and +4 actinide ions exist as discrete ions which are hydrated in solution as

An(H2O)n+3/+4

species. The +5 and +6 oxidation states however hydrolyze water in aqueous

solutions to form actinyl, AnO2+ and AnO2

2+ ions respectively. The An

+3/+4 and AnO2

+/2+ ions

form a large number of coordination complexes by binding various types of ligands. This is due

to the hard acid nature of these ions. They form strong complexes with hard bases such as

anionic or oxygen donating ligands. Examples of such complexes are thorium nitrate

(Th(NO3)4.5H2O), actinide halides (AnX4, An = Th, U, Np and Pu, X = F, Cl and Br) and

uranium carbonate, UO2(CO3)34-

. Complexes formed with organic ligands are also well known.

The overall coordination number of the An+3/+4

and AnO2+/2+

ions depend on the size and shape

(steric effect) of the ligands. For example, the effective coordination numbers for aquo ligands to

these ions are 8-10 and 9-12 for the trivalent and tetravalent ions respectively and 4-5 and 5-6,

especially for the aquo ligand, for the pentavalent and hexavalent actinyl species respectively.

The aquo ions for U(VI) and U(IV) with coordination numbers of 5 and 8 respectively are shown

in Figure 1.2.

As uranium is the most abundant and most widely used actinide element, a larger

(compared to the other actinides) proportion of experimental studies have focused in

synthesizing and characterizing its complexes. Although there has been recent progress in the

synthesis and characterization of diverse uranium complexes, there are still a lot of experimental

Page 30: Relativistic Quantum Chemistry Applied to Actinides

8

challenges to our understanding of the behavior of these elements and their compounds in the

environment. A few of them are: 1) the need for difficult handling techniques as a result of the

Figure 1.2: The aquo complexes of U(VI) and U(IV).

toxicity and radioactivity of the actinide elements and the relative scarcity of the actinide

elements. This is especially true for the trans-uranium elements. In many cases, several different

oxidation states might be simultaneously present in solution further complicating the work of

experimental chemists. In contrast, computational studies are relatively cheap, safe and can be

used to either complement available experimental data or bridge any gaps in our understanding

of the properties of actinide complexes.

Page 31: Relativistic Quantum Chemistry Applied to Actinides

9

1.3 Theoretical studies of actinide complexes.

1.3.1 The Schrödinger equation

All theoretical approaches employed in this thesis begin with a formulation of the time-

independent adiabatic approximation of the Schrödinger equation.13

In these methods, the

Schrödinger equation is essentially recast to take advantage of the Born-Oppenheimer (BO)

approximation14

which allows the motion of the nuclei and electrons to be uncoupled. The

physical meaning of the BO approximation is an assumption that the nuclei moves at negligible

speeds compared to the electrons. The nuclei are assumed to be stationary in comparison to the

fast moving electrons. This separation of the nuclear and electronic motion allows for an easier

evaluation of the electron-nuclei, nuclei-nuclei interactions as well as the nuclear kinetic

energies.

(1.1)

(1.2)

(1.3)

(1.4)

Vnuclei is constant at each molecular geometry (BO approximation) and so can be

removed. This allows us to recast Equation 1.1 into the electronic Schrodinger equation,

Equation 1.3. The variational minimization of the energy results in the ground state

wavefunction, Ψ and the lowest electronic energy, E, obtained during the variational

minimization depends parametrically on the nuclear positions. The components of Equation 1.4

are the kinetic energy of the electrons, the electron-nuclei interaction potential and the electron-

Page 32: Relativistic Quantum Chemistry Applied to Actinides

10

electron interaction potential. The minimization of the energy, complete solution of the

Schrödinger equation, is however not tractable except for the smallest systems such as H2, H2+,

and He+. The reason for this is the 3N-dimensional nature of the electronic wavefunction, where

N is the number of electrons. The electron-electron interaction part, , of the

Hamiltonian is also very difficult to evaluate.

1.3.2 The variational principle and electronic basis sets.

Examination of Equation 1.3 shows us that it has an infinite number of solutions.

Assuming we are working with the electronic Schrodinger equation, the electronic energy

obtained with an arbitrary wavefunction can be written as in Equation 1.5. The solution that

yields the lower bound to the energy, Eel, yields the correct ground state wave-function. The

lower bound to the energy is labeled as E0. The variational principle in lay-man terms simply

states “if there are two wavefunctions for a system, the one that produces the lower energy better

represents the ground state wavefunction of the system and in fact the true ground state

wavefunction yields the lowest energy, E0”. According to the variational principle, if the exact

ground state wavefunction is used in Equation 1.5, the lowest electronic energy will be obtained.

Thus Equation 1.5 can be generalized to Equation 1.6.

Page 33: Relativistic Quantum Chemistry Applied to Actinides

11

It is immediately obvious that to get a given value of Eel as described in Equation 1.6, we

need a trial wavefunction, its quality (seen as Eel – E0) notwithstanding. For atomic and

molecular systems, electronic trial wavefunctions need to fulfill two important properties. Firstly

as electrons are fermions, the trial wavefunction must be anti-symmetric under particle

interchange.

And secondly, no two identical electrons (or fermions in general) can occupy the same quantum

state simultaneously. This is called the Pauli Exclusion Principle. The Slater determinant is an

expression that conforms to these two conditions for multielectronic systems.15

It is written as a

determinant consisting of several orthonormal spin-orbitals each of which describe the position

and spin of an electron.

There are generally two classes of functions used to represent atomic spin-orbitals in

modern electronic structure theory.16

The first are called Slater type orbitals (STOs) which are

modeled using the exponentially decaying electronic distribution (e-ζr

) of the hydrogen atom. As

they correspond to a real atom, they accurately describe the cusp condition at the atomic center.

There are however significant computational challenges to calculating the integrals of an

exponential decay function. In most cases, numerical solutions to the integrals of these functions

are needed. The cusp at the atomic center is also difficult to handle computationally. The

Gaussian type orbitals (GTOs) are a more tractable class of orbitals. Such orbitals possess

Page 34: Relativistic Quantum Chemistry Applied to Actinides

12

analytical solutions to their integrals. The disadvantages of using GTOs are: 1) wrong cusp

conditions and 2) overall poor long range behavior as they decay much faster than STOs, e-αr2

instead of e-ζr

. A widely used approach to circumvent these disadvantages is to use several GTOs

to represent a single basis function according to Equation 1.9. The coefficients, Ci, are fit to

ensure agreement with the radial electronic distribution of the hydrogen atom.

In molecular systems, the atomic basis sets are combined to represent a molecular orbital

(MO) as shown in Equation 1.10. This is called the Linear Combination of Atomic Orbitals

principle (LCAO). The coefficients, Si, describe the contribution of each atomic orbital to the

MO. Prior to going forward, we note that the constituent atomic orbitals of an MO can contain

one basis function (either an STO or GTO). These types of basis sets are said to be of single-ζ

quality. Those containing two, three and four basis functions per atomic orbital are said to be of

double-ζ, triple-ζ and quadruple-ζ basis sets. Polarization functions (basis functions with higher

angular momentum) as well as diffuse basis functions can be added to each atomic orbital to

respectively allow for a better description of electron correlation and anionic systems or excited

electronic states. In some cases, the basis function is partitioned into a core and valence part.

This is generally rooted in the idea that core electrons are generally chemically inactive and not

involved in bond formation or ionization processes. A more complete explanation of the

variational principle and the other concepts briefly described here can be found in most modern

computational chemistry textbooks.16-17

Page 35: Relativistic Quantum Chemistry Applied to Actinides

13

1.3.3 The Hartree-Fock method and Post Hartree-Fock approaches.

At the Hartree-Fock level, a Fock operator consisting of one-electron kinetic energy, the

nucleus-electron interaction potential and the Hartree-Fock potential is defined, Equation 1.11.

The Hartree-Fock method is the direct result of applying the variational method to the Slater

determinant, Equation 1.8, in the electronic Schrödinger equation, Equation 1.1. The first two

terms of Equation 1.11 (the electron kinetic energies and nucleus-electron interaction potentials)

are one-electron operators describing the motion of an ith

electron in the field of the nuclei. These

are combined into the term of Equation 1.12. The Hartree-Fock potential consists of the

Coulomb (Ĵj) and exchange (Ǩ j) operators. It is a two-electron operator that describes inter-

electron repulsion. The former, Ĵj, describes the classical Coulombic repulsion between the

electron and the jth

electron. The exchange operator, Ǩ j, defines the electron exchange energy

and it essentially switches the spin orbital of the ith

electron with that of the jth

electron. The

Hartree-Fock Hamiltonian is then described as the sum of the Fock operators for all electrons in

the system, Equation 1.14.18-19

This equation allows us to write the Hartree-Fock (HF) energy as

Page 36: Relativistic Quantum Chemistry Applied to Actinides

14

It becomes immediately obvious that HF theory, from the definition of the Hartree-Fock

potential, is a mean-field approximation in which each electron moves in an average field

generated by the remaining (n-1) electrons. There are two very important deficiencies to the HF

solution with a single Slater determinant. The first relates to the difference between the

instantaneous electron-electron interaction of the real system and the mean field interaction

experienced by each electron in HF theory. This is called dynamical correlation. The second

deficiency is labeled as non-dynamical correlation and relates to deficiencies caused by the use

of a single Slater determinant to describe the system. Non-dynamical correlation is particularly

large for systems with near-degenerate energy levels. As such the true wavefunction of such

systems consist of several coefficient-weighted Slater determinants. The sum of the dynamical

and non-dynamical correlation is called the correlation energy and is defined as the difference

between the ground state electronic energy and the HF energy.

It is important to note that although the contribution of electron correlation to the total

electronic energy is small (~1%), its omission in most cases leads to large errors in calculated

structural and electronic properties as well as reaction energies. There are various approaches to

improving the HF approach. The only post Hartree-Fock approaches employed in this work are

the second-order Møller-Plesset perturbation theory, MP2 and the coupled cluster singles and

doubles and perturbatively included triples, CCSD(T), approaches.

Page 37: Relativistic Quantum Chemistry Applied to Actinides

15

In Møller-Plesset perturbation theory,20-21

the unperturbed Hamiltonian, , is the sum of

the one-electron Fock operators in Equation 1.13, and the Hartree-Fock wavefunction is an

eigenfunction of Ĥ0 which yields an eigenvalue equal to the sum of the one electron energies of

the occupied spin orbitals. The electron correlation (dynamical correlation as the single Slater

determinant is still employed) is described as a perturbation, Ĥ′, to Ĥ

0.

Essentially a subset of the general time-independent perturbation theory of Rayleigh and

Schrödinger22

, Møller-Plesset perturbation theory assumes that all Slater determinants

corresponding to the excitation of electrons from the occupied to the virtual orbitals are also

eigenfunctions of Ĥ0 with an eigenvalue equal to the sum of the one electron energies of their

occupied spin orbitals. The wavefunction and energy can be expanded in terms of the

perturbation.

Substitution of these equations into the electronic Schrödinger equation yields a series of

equations conforming to the general formula: and .

The sum of and describes the HF energy (combination of the one-electron energy and the

Hartree-Fock potential). The various Møller-Plesset perturbation (MP) approaches are named

according to the degree of correlation correction included with the HF potential. For example, in

MP2, is included. Lastly, it should be noted that the higher orders of the

MPn approaches are not necessarily convergent with respect to total energies.

16

Page 38: Relativistic Quantum Chemistry Applied to Actinides

16

In coupled cluster theory, an infinite exponential cluster operator acts on the Hartree-

Fock wavefunction to generate a full configuration interaction, CI, solution accounting for

electron correlation, Equation 1.19.23

The coupled cluster approaches are labeled according to

the level at which the expansion coefficients, Ť, are truncated, Equation 1.20.24

As an example,

truncation at the single and double excitations level leads to the coupled cluster singles and

doubles approach (CCSD). The most popular coupled cluster approach involves the inclusion of

triple excitations into the CCSD wavefunction in a perturbative (Møller-Plesset) manner,

Equation 1.16. This method is called the CCSD(T) approach. The CCSD(T) method is often

referred to as the gold standard of computational chemistry as it has been shown to yield highly

accurate (< 1 kcal/mol) reaction energies, transition state barriers and structure.24-25

It is also a

compromise between computational expense and the more accurate CCSDTQ and higher order

coupled cluster approaches.

To briefly review, the major deficiency of HF theory is the lack of an appropriate

treatment of the instantaneous interaction between electrons as well as the single reference nature

of the Slater determinant employed. This neglect of electron correlation and mean field

approximation to inter-electron repulsion has dramatic effects on calculated bond energies,

vibrational frequencies and bond lengths in molecular species.16

The MPn and coupled cluster

approaches become increasingly expensive at higher orders. The energy convergence of the MPn

methods is in many cases oscillatory at increasingly higher orders. These approaches also

experience significant spin contaminations when employed for open shell (unrestricted HF

Page 39: Relativistic Quantum Chemistry Applied to Actinides

17

wavefunction) systems. The coupled cluster techniques are only computationally feasible (at the

current moment) for the smallest of molecules. They also require basis sets of at least triple-ζ

quality.

1.3.4 Density Functional Theory

Most of the calculations in this thesis were carried out under the framework of density

functional theory (DFT). Although it has its roots in the much earlier Thomas-Fermi model26-27

and the work of Slater on the Xα exchange functional15, 28

, modern DFT is based on the two

fundamental Hohenberg-Kohn (HK) theorems.29

The first HK theorem states that the ground

state electron density, ρ(r), uniquely determines the external potential V(r). The determination of

V(r) allows for the exact formulation of the Hamiltonian and thus the wavefunction. The second

HK theorem states that the ground state energy can be obtained variationally. This is due to the

fact that any new density generates a new external potential leading to a new wavefunction. The

energy, a functional of the density, is the sum of the external potential, , the kinetic

energy, and electron-electron interaction energy, , Equation 1.21. The

Eee term contains both the classical Coulombic interactions and the non-classical electron-

electron interactions.

Kohn-Sham DFT is the most widely used form of DFT. The basic assumption to this

framework is the stipulated existence of a system of non-interacting electrons with exactly the

Page 40: Relativistic Quantum Chemistry Applied to Actinides

18

same density as the system of interest, one with interacting electrons.30

For the system of non-

interacting electrons, the Eee term of Equation 1.21 by definition becomes zero. An analogous

equation to the Fock equations of HF theory (Equations 1.11-1.14) allows us to construct Kohn-

Sham spin orbitals. in Equation 1.23 is the energy of the spin orbitals generated for the

non-interacting system. Equation 1.24, is an eigen-value equation for Kohn-Sham particles and is

reminiscent of the Fock equations at the Hartree-Fock level, Equation 1.12.

The one-electron Hamiltonian for the Kohn-Sham particles is given as

The effective external potential is defined as

The energy of the Kohn-Sham system is then given as

The exchange-correlation term, , is a composite term containing: 1) the kinetic energy

of a system of interacting electrons minus the kinetic energy of a system with non-interacting

electrons with exactly the same density and 2) the electron-electron interaction (total electron

exchange and correlation) minus the classical Coulomb interaction. Working backwards, we can

Page 41: Relativistic Quantum Chemistry Applied to Actinides

19

see that the Schrödinger equation has essentially been recast with energy as functional of density.

If the exact exchange-correlation term is known, a full solution to the Schrödinger equation will

have been obtained. Although a large number of possible exchange-correlation functionals have

been proposed and tested over the years,31-36

the exact nature of this functional remains

unknown. This functional is often split into the exchange functional and the correlation

functional.

The earliest known exchange-correlation functional is called the local density

approximation (LDA) and is based on the fictitious system of a uniform electron gas, UEG. This

system has an infinite number of electrons and uniform density all throughout. The exchange

energy of a UEG is a functional of its density, Equation 1.29. The correlation part of the

exchange-correlation functional was parameterized using highly accurate Monte Carlo

calculations on the UEG. This work was done by Ceperley and Alder.37

Modern modifications

such as that of Vosko, Wilk and Nusair are available in most modern ab initio software

packages.38

Testing of the performance of the LDA functional for actinide molecules reveals that

it generally leads to bond lengths that are slightly shorter than experimental values as well as

reaction energies that deviate significantly from the experimental values. LDA is however

particularly suited for metallic periodic systems where the electron density changes only very

slowly.16-17

Page 42: Relativistic Quantum Chemistry Applied to Actinides

20

The most common modifications to the LDA functional involve attempts to correct for

rapidly changing electron densities found in molecular systems. The generalized gradient

approximation, GGA, functionals were the first corrections to find widespread usage for

molecular systems. In these functionals the gradient of the density is included in the formulation

of the exchange-correlation functional, Equation 1.30. The term is the gradient

parameter and provides a better description of the electron exchange in regions where there are

changes in the electron density. The GGA functionals provided vast improvement on the

agreement between the calculated ionization, atomization and binding energies and the

experimental values.16

Rigorous tests have however found that they generally tend to slightly

overestimate the lengths of the bonds in actinide complexes.39-43

The PBE functional is a non-

empirical GGA functional and was widely used in the works compiled in this thesis.44-45

The next stages of modification to the exchange-correlation functional encompass the

hybrid and meta-GGA functionals. Additional information regarding the density is included as

the Laplacian of the density and the non-interacting kinetic energy in the meta-GGA functionals.

No meta-GGA functionals were employed in the works compiled in this thesis. Regarding the

hybrid functionals, the gradient corrected GGA functionals are combined with explicit Hartree-

Fock exchange. An example of the general mixing of GGAs with Hartree-Fock exchange as seen

in the Becke three parameter functionals is shown in Equation 1.31. This combination is based

on the adiabatic connection formula which allows us to write the exchange-correlation formula

as a combination of the Hartree-Fock exchange and some DFT exchange-correlation functional.

Page 43: Relativistic Quantum Chemistry Applied to Actinides

21

The most widely used functional in the works compiled in this thesis and all of computational

actinide chemistry in general is the B3LYP functional.46-47

It combines some portion of the

Hartree exchange with the B88 exchange and LYP correlation functionals using the formula:

The B88 exchange functional of Becke and the popular LYP correlation functional of Lee Yang

and Parr are the GGA functionals employed in the B3LYP functional. Extensive testing by

various authors have shown this particular functional to be suited for calculating the structural

properties of actinide complexes, their vibrational frequencies as well as their reaction

energies.39, 41-43, 48

Other examples of hybrid functionals include B3PW91 and BHandH.16

Other functionals employed in this work are the long-range corrected hybrid functional,

CAM-B3LYP49

, the half and half functional of Becke, BHandH46

, and the B3LYP functional

with dispersion corrections using Grimme‟s third scheme, B3LYP-D3.33

1.3.5 Relativistic effects

The time-independent Schrödinger equation13, 22

is only valid for non-relativistic systems.

As the atomic number, Z, increases in heavier nuclei, the speeds of the core-electrons approach

the speed of light. At such speeds, relativistic effects become important in the electronic structure

of the atoms, ions and compounds of heavy nuclei. Generally the major effects in molecular

systems are propagated by the contraction and stabilization of s and p orbitals as well as the

expansion and destabilization of d and f orbitals.50-52

Relativistic effects increase according to

Z2/c

2 down the periodic table, where c is the speed of light, 137 in atomic units. The Dirac

equation, Equation 1.32, was formulated in 1929 to describe the one-electron energies of

relativistic systems53

and is similar to the earlier Klein-Gordon equation54

. In the Dirac equation,

Page 44: Relativistic Quantum Chemistry Applied to Actinides

22

α and β are 4×4 matrices, Equations 1.33-1.35, V is the external potential and the wavefunction is

a four component column spinor, Equation 1.36, containing the large component, ΨL, and the

small component, ΨS. The large and small components respectively describe the positive and

negative energy solutions found on either side of 2mc2. The small component solutions

correspond to positrons while the large component solutions correspond to electrons. The up or

down arrows in Equation 1.36 indicate the particle spin of the electrons and positrons. The

coupling between the positronic and electronic parts of the wavefunction can be removed

through the Foldy-Wouthuysen (FW) transformation.55

The Pauli and Zeroth order regular

approximation (ZORA) Hamiltonians were obtained by first order FW transformation of the

Dirac-Coulomb-Breit Hamiltonian expanded in a series of c-2

and E/(2mc2-V).

56-57 Calculations

in which the spin components are projected out of the four-component Dirac equation are known

as scalar-relativistic calculations while those in which the small-component is projected out are

two-component approximations to the full Dirac equation. In these sets of equations (Equations

1.32-1.36), I is the two-dimensional identity matrix, σ represents the Pauli spin matrices in a

compact form and p is the momentum.

Page 45: Relativistic Quantum Chemistry Applied to Actinides

23

Modern quantum-chemical calculations generally include relativistic effects using any

one of the ZORA56-57

, the Douglas-Kroll-Hess58-59

unitary transformation and relativistic

effective core potentials (ECP)60-61

. The ZORA and ECP approaches were employed in the

calculations summarized in this thesis. We however here focus on discussing the use of

relativistic ECPs as they were used in nearly all the studies included in this work.

The realization that the core electrons unlike their valence counterparts are chemically

inert and not involved in bond formation implies that they can be completely frozen (frozen-core

approximation) or replaced by an effective core potential (ECP) without any massive loss in

accuracy. The overall Hamiltonian used in the ECP calculations is transformed into Equation

1.37 below. In this equation, contains the Coulomb and exchange operators for the

valence electrons and potential to account for the core electrons. The kinetic and electron-

electron correlation terms are completely non-relativistic.

The design of ECPs usually starts from all-electron basis set calculations on an atom of

the desired element. The design could either be for non-relativistic ECPs in which case an HF

(and in some cases post-HF) wave-function is employed or for relativistic ECPs in which some

approximation to the Dirac equation (in most cases, the quasi relativistic Pauli Hamiltonian) is

used. For the latter, all relativistic effects are embedded in the Vpp term of Equation 1.37. The

designers of either flavor of ECPs however face certain challenges:62-63

A) Would the ECP be

Page 46: Relativistic Quantum Chemistry Applied to Actinides

24

constructed such that it replicates the shapes and energies and shapes of the orbitals (shape-

consistent ECPs) or would it be an energy-consistent ECP which would replicate some electronic

property (ionization potentials, electron affinities or electronic spectra) of the atom? The energy-

consistent ECPs are the most widely used variety as a result of the availability of experimental

data regarding the electronic properties of most atoms. B) Where is the core-valence boundary?

In other words, which orbitals should be placed in the core region to be replaced by the ECP and

which should be in the valence region? For the actinide elements, this particular question is non-

trivial given the semi-core nature of the 6s and 6p orbitals as well as the participation of the 5f

orbitals in bond formation.

There are two commonly used types of energy-consistent ECPs used in computational

studies of actinide systems.64-66

The design of the small-core (SC) ECPs is such that 60 (principal

quantum number, n < 5 shells) core electrons are represented with a pseudopotential while the

remaining (n ≥ 5 shells) electrons (30 for thorium, 32 for uranium as examples) are represented

by valence basis sets. For the second variety, the large-core (LC) ECPs, 78 core electrons are

replaced by a pseudopotential. As the core-region covered by the pseudopotential is larger for the

LC-ECPs, they provide even greater time savings than the SC-ECPs. However, it has been

shown that SC-ECPs generally tend to provide better agreements between experimental and

calculated structural parameters and reaction energies. In Chapter 2 and works by other authors,

the performance of the SC-ECPs and LC-ECPs designed by Stuttgart-Cologne group relative to

experimental results (of the structure and reaction energies of several uranium complexes) are

compared to those obtained using a full scalar relativistic four-component approximation to the

Dirac equation.41-43

Page 47: Relativistic Quantum Chemistry Applied to Actinides

25

1.3.6 Solvation Effects

The effect of a solvent environment on the nature, speciation/chemical form and spectra

of actinide complexes can be significant. For example for the meta-stable U(V) complexes, an

aqueous solvent is sufficiently oxidizing to prevent their isolation and characterization.12, 67-71

Experimental NMR studies of UO22+

in alkaline solution have shown the existence of fast

oxygen exchange processes.48, 72-74

It should be noted that the behaviors of actinide species in

solutions are in no way uniform. For example, a comparison of the speciation diagrams of UO22+

and PuO22+

, indicated that although the pentaaquo complex is dominant for both species in

highly acidic solutions, the uranyl moiety forms trinuclear (in addition to binuclear) species,

[(UO2)3(OH)5]+ and [(UO2)3(OH)7]

2- at modest to high pH values while its plutonyl counterpart

forms mainly the binuclear species, [(PuO2)2(OH)2]2+

with negligible concentrations of trinuclear

plutonyl complexes, even at high pH values.75-77

The inclusion of solvent effects, when needed,

is therefore very important in modern computational actinide chemistry.

One approach for describing the effect of a solvent environment on the structure,

speciation and electronic properties of actinide complexes involves adding a large number of

solvent molecules around the solute molecule. As an example, to study the uranyl ion in aqueous

solution, the UO22+

ion could be surrounded by a cubic box containing water molecules with an

overall density of about 1.0 g/cm3. Vibrational frequency analyses and other calculations are then

carried out on this solute-solvent box after the structural optimizations. This explicit approach to

modeling solvation effects can be extremely computationally demanding as the number of

solvent molecules increase. In addition, the introduction of so many solvent molecules implies

the existence of a large number of possible minima and transition state structures. This further

complicates any search for the global minimum structure.16

Page 48: Relativistic Quantum Chemistry Applied to Actinides

26

For the implicit solvation models,78-82

a polarizable continuum or conductor-like model

with electrostatic and entropic properties that match that of the desired solvent is employed. For

these solvation models, the size/shape of the solvent-excluded cavity around the solute molecule

has to be defined. The molecular free energy in solution is calculated as a sum of the

electrostatic, dispersion-repulsion and cavitation energy contributions, Equation 1.38.

The cavity can be described such that each atom in a molecule is represented by

individual spheres, individual cavity models, or for the united atom models, in which the

hydrogen atoms are included in the spheres of the atoms to which they are directly bonded. Most

software suites allow the user to specify the formation of the cavity from various tesserae as well

as the radii of each cavity. The implementation of the polarizable continuum solvation model83-84

(PCM) in the Gaussian 03 suites of programs85

was used in most of the work compiled in this

thesis. Gutowski et al.86-87

have previously explored the suitability of this model and other

solvation models for describing the effect of solvent environments on the structures and

stabilities of actinide complexes. It has been shown that implicit model calculations on AnO22+

and AnO2+ ions with a first solvation sphere containing 4-6 explicit water molecules accurately

reproduces the structure and properties of these ions in highly acidic aqueous solutions.41-42, 88

1.4 Organization of this Thesis

This thesis is written in a sandwich style agglomeration of several manuscripts published

or submitted in peer-reviewed scientific journals during the course of the doctoral program. The

over-arching aim of the dissertation is to further our understanding of actinide chemistry by

using computational methods. In each chapter, we try to answer particular questions regarding

Page 49: Relativistic Quantum Chemistry Applied to Actinides

27

the structure and properties of actinide complexes. The structural and electronic properties as

well as the chemistry of actinide species in the gaseous, aqueous and solid phases are examined

using theoretical calculations. In addition, the performances of the various theoretical methods

used were also examined. Each chapter is followed by a list of references. The basic outline of

this thesis is illustrated in Figure 1.3.

A brief introduction to actinide chemistry and the computational methods used in this

thesis are presented in Chapter 1. In Chapter 2, we attempted to benchmark the performance of

LC-ECP and SC-ECPs against an all-electron basis set four-component scalar-relativistic

approach in the framework of DFT. This chapter essentially answers the question: how much

accuracy can be obtained with ECP calculations? And how accurate are DFT calculations

employing relativistic ECPs compared to those employing all-electron electron basis sets with a

four-component approximation to the Dirac Hamiltonian?

In Chapters 3, 4 and 5, several questions regarding the aqueous chemistry of actinide

complexes were studied using DFT calculations. The ability of DFT, a single-reference theory to

accurately describe the structure of plutonium complexes is confirmed in Chapter 3. It is very

important to ensure that the structures of open-shell plutonium complexes can be well-predicted

by the single reference Kohn-Sham DFT approach before using the approach in studying their

speciation and energetics. Given that we found DFT to be well suited to predicting the structures

of plutonyl (VI) complexes in Chapter 3, we proceeded to studying the structural properties,

speciation and energetics of the PuO22+

in acidic and alkaline aqueous solutions in Chapter 4.

This study allowed us to explain the differences between the trinuclear species-dominated uranyl

chemistry at modest-high pH and the binuclear species-dominated plutonyl chemistry at similar

pH values. In Chapter 5, DFT calculations were used to explore the structure and speciation of

Page 50: Relativistic Quantum Chemistry Applied to Actinides

28

Figure 1.3: Schematic description of the chapters in this thesis

mononuclear uranyl peroxo complexes in aqueous solution. These complexes are possibly the

building blocks of the well-known crystalline polynuclear uranyl-peroxide species.

In the next chapter, Chapter 6, the adsorption, electronic structure and reactions of

uranium complexes on geochemical surfaces were studied using a combination of periodic and

molecular DFT calculations. A comprehensive study of the structure and bonding as well as the

electronic properties of uranyl fluorides in the gas-phase, aqueous phase and in the cavities of a

protein is presented in Chapter 7. This study encompasses three of the phases (except for the

solid phase) in which actinides in the environment can be found.

The next two chapters (Chapters 8 and 9) concentrate on novel uranium chemistry. As

previously noted8, the U(V) ion is metastable as it disproportionates to U(VI) and U(IV). New

Page 51: Relativistic Quantum Chemistry Applied to Actinides

29

experimental breakthroughs have however led to the synthesis of stable U(V) species. The

structural and electronic properties of two binuclear U(V)/U(V) complexes formed by oxo-

functionalization of axial oxo atoms are studied in Chapter 8. In Chapter 9 however, the structure

of a binuclear U(VI)/U(VI) complex formed by laser ablation of UO3 is studied. Particular

attention was given to possible structural motifs featuring cation-cation interactions.

In Chapter 10, the results from the various studies compiled in this thesis are presented.

The linkages between the different chapters and how each chapter relates to the goal of effective

nuclear waste storage and environmental remediation of nuclear waste sites are discussed.

Finally, future directions for the works compiled in this thesis as well as for computational

actinide chemistry in general are discussed.

1.5. References

1. Kaltsoyannis, N.; Scott, P., The f elements,. 1 ed.; Oxford University Press: Oxford,

1999.

2. Cotton, F. A.; Wilkinson, G., Advanced Inorganic Chemistry. 1 ed.; John Wiley and

Sons: New York, 1988.

3. Eliav, E.; Kaldor, U.; Ishikawa, Y., Phys. Rev. A 1995, 52 (1), 291-296.

4. http://www.mindat.org/. (Assessed July 23rd

, 2012)

5. Holmes, D. http://californiahomeopath.com/issue/the-california-homeopath-volume-14-

1/article/the-actinides-the-ultimate-challenge. (Assessed July 23rd

, 2012)

6. http://www.world-nuclear.org/info/inf01.html. (Assessed July 23rd

, 2012)

7. Ewing, R. C.; Runde, W.; Albrecht-Schmitt, T. E., Mrs Bulletin 2010, 35 (11), 859-866.

Page 52: Relativistic Quantum Chemistry Applied to Actinides

30

8. Runde, W. The Chemical Interactions of Actinides in the Environment Los Alamos

Science [Online], 2000.

9. Betti, M., J. Environ. Radioactiv. 2003, 64 (2-3), 113-119.

10. Clain, A. F.; de Aquino, J. O.; Domingues, M. D. F., Quim. Nova 1999, 22 (5), 677-678.

11. https://www.orau.org/ptp/collection/Miscellaneous/pacemaker.htm. (Assessed July 23rd

,

2012)

12. Graves, C. R.; Kiplinger, J. L., Chem. Commun. 2009, (26), 3831-3853.

13. Schrodinger, E., Annalen Der Physik 1926, 79 (4), 361-U8.

14. Born, M.; Oppenheimer, R., Annalen Der Physik 1927, 84 (20), 0457-0484.

15. Slater, J. C., Phys. Rev. 1951, 81 (3), 385-390.

16. Cramer, C. J., Essentials of Computational Chemistry -Theories and Models. 2 ed.; John

Wiley & Sons Ltd: London, 2005.

17. Jensen, F., Introduction to Computational Chemistry. Wiley: Chichester, 1999.

18. Hartree, D. R., P. Camb. Philos. Soc. 1928, 24, 89-110.

19. Hartree, D. R., P. Camb. Philos. Soc. 1928, 24, 111-132.

20. Moller, C.; Plesset, M. S., Phys. Rev. 1934, 46 (7), 0618-0622.

21. Pople, J. A.; Binkley, J. S.; Seeger, R., Int. J. Quantum Chem. 1976, 1-19.

22. Schrodinger, E., Annalen Der Physik 1926, 80 (13), 437-490.

23. Cizek, J., J. Chem. Phys. 1966, 45 (11), 4256-&.

24. Purvis, G. D.; Bartlett, R. J., J. Chem. Phys. 1982, 76 (4), 1910-1918.

25. Kendall, R. A.; Dunning, T. H.; Harrison, R. J., J. Chem. Phys. 1992, 96 (9), 6796-6806.

26. Thomas, L. H., P. Camb. Philos. Soc. 1927, 23, 542-548.

27. Fermi, E., Rend. Accad. Lincei 1927, 6, 602-607.

Page 53: Relativistic Quantum Chemistry Applied to Actinides

31

28. Slater, J. C., Adv. Quantum Chem. 1972, 6, 1-.

29. Hohenberg, P.; Kohn, W., Phys. Rev. B 1964, 136 (3B), B864-&.

30. Kohn, W.; Sham, L. J., Phys. Rev. 1965, 140 (4A), 1133-&.

31. Cohen, A. J.; Mori-Sanchez, P.; Yang, W. T., Chem. Rev. 2012, 112 (1), 289-320.

32. Cramer, C. J.; Truhlar, D. G., Phys. Chem. Chem. Phys. 2009, 11 (46), 10757-10816.

33. Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H., J. Chem. Phys. 2010, 132 (15).

34. Kozuch, S.; Gruzman, D.; Martin, J. M. L., J. Phys. Chem. C 2010, 114 (48), 20801-

20808.

35. Zhao, Y.; Truhlar, D. G., Chem. Phys. Lett. 2011, 502 (1-3), 1-13.

36. Zhao, Y.; Truhlar, D. G., J. Chem. Theory Comput. 2011, 7 (3), 669-676.

37. Ceperley, D. M.; Alder, B. J., Phys. Rev. Lett. 1980, 45 (7), 566-569.

38. Vosko, S. H.; Wilk, L.; Nusair, M., Can. J. Phys. 1980, 58 (8), 1200-1211.

39. Batista, E. R.; Martin, R. L.; Hay, P. J., J. Chem. Phys. 2004, 121 (22), 11104-11111.

40. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2011, 115 (48), 14110–14119.

41. Schreckenbach, G.; Shamov, G. A., Acc. Chem. Res. 2010, 43 (1), 19-29.

42. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109 (48), 10961-10974.

43. Shamov, G. A.; Schreckenbach, G.; Vo, T. N., Chem-Eur. J. 2007, 13 (17), 4932-4947.

44. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1996, 77 (18), 3865-3868.

45. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1997, 78 (7), 1396-1396.

46. Becke, A. D., J. Chem. Phys. 1993, 98 (7), 5648-5652.

47. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J., J. Phys. Chem. 1994, 98

(45), 11623-11627.

48. Schreckenbach, G.; Hay, P. J.; Martin, R. L., Inorg. Chem. 1998, 37 (17), 4442-4451.

Page 54: Relativistic Quantum Chemistry Applied to Actinides

32

49. Yanai, T.; Tew, D. P.; Handy, N. C., Chem. Phys. Lett. 2004, 393 (1-3), 51-57.

50. Pyykko, P., Chem. Rev. 1988, 88 (3), 563-594.

51. Pyykko, P.; Desclaux, J. P., Acc. Chem. Res. 1979, 12 (8), 276-281.

52. Pyykko, P.; Li, J.; Runeberg, N., J. Phys. Chem. 1994, 98 (18), 4809-4813.

53. Dirac, P. A. M., Proc. Roy. Soc. Ser. A, 1929, 123, 714. 1929, 123, 714-.

54. Polyanin, A. D., Handbook of Linear Partial Differential Equations for Engineers and

Scientists. Chapman & Hall/CRC: 2002.

55. Foldy, L. L.; Wouthuysen, S. A., Phys. Rev. 1950, 78 (1), 29-36.

56. Dyall, K. G.; van Lenthe, E., J. Chem. Phys. 1999, 111 (4), 1366-1372.

57. van Lenthe, E.; Ehlers, A.; Baerends, E. J., J. Chem. Phys. 1999, 110 (18), 8943-8953.

58. Reiher, M.; Wolf, A., J. Chem. Phys. 2004, 121 (5), 2037-2047.

59. Wolf, A.; Reiher, M.; Hess, B. A., J. Chem. Phys. 2002, 117 (20), 9215-9226.

60. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74 (6), 1245-1263.

61. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100 (10), 7535-7542.

62. Dolg, M.; Cao, X. Y., Chem. Rev. 2012, 112 (1), 403-480.

63. Stoll, H.; Metz, B.; Dolg, M., J. Comput. Chem. 2002, 23 (8), 767-778.

64. Cao, X. Y.; Dolg, M., J. Mol. Struc-Theochem 2004, 673 (1-3), 203-209.

65. Cao, X. Y.; Dolg, M., Coordin. Chem. Rev. 2006, 250 (7-8), 900-910.

66. Cao, X. Y.; Dolg, M.; Stoll, H., J. Chem. Phys. 2003, 118 (2), 487-496.

67. Arnold, P. L.; Pecharman, A. F.; Love, J. B., Angew. Chem. Int. Edit. 2011, 50 (40),

9456-9458.

68. Fortier, S.; Hayton, T. W., Coordin. Chem. Rev. 2010, 254 (3-4), 197-214.

Page 55: Relativistic Quantum Chemistry Applied to Actinides

33

69. Ikeda, A.; Hennig, C.; Tsushima, S.; Takao, K.; Ikeda, Y.; Scheinost, A. C.; Bernhard,

G., Inorg. Chem. 2007, 46 (10), 4212-4219.

70. Mizuoka, K.; Grenthe, I.; Ikeda, Y., Inorg. Chem. 2005, 44 (13), 4472-4474.

71. Mizuoka, K.; Ikeda, Y., Inorg. Chem. 2003, 42 (11), 3396-3398.

72. Clark, D. L.; Conradson, S. D.; Donohoe, R. J.; Keogh, D. W.; Morris, D. E.; Palmer, P.

D.; Rogers, R. D.; Tait, C. D., Inorg. Chem. 1999, 38 (7), 1456-1466.

73. Shamov, G. A.; Schreckenbach, G., J. Am. Chem. Soc. 2008, 130 (41), 13735-13744.

74. Szabo, Z.; Grenthe, I., Inorg. Chem. 2007, 46 (22), 9372-9378.

75. Rao, L. F.; Tian, G. X.; Di Bernardo, P.; Zanonato, P., Chem-Eur. J. 2011, 17 (39),

10985-10993.

76. Reilly, S. D.; Neu, M. P., Inorg. Chem. 2006, 45 (4), 1839-1846.

77. Zanonato, P.; Di Bernardo, P.; Bismondo, A.; Liu, G. K.; Chen, X. Y.; Rao, L. F., J. Am.

Chem. Soc. 2004, 126 (17), 5515-5522.

78. http://www.nd.edu/~wschnei1/courses/CBE_547/Lectures/implicit_solvation_models.pdf.

(Assessed July 23rd

, 2012)

79. Cramer, C. J.; Truhlar, D. G., Chem. Rev. 1999, 99 (8), 2161-2200.

80. Cramer, C. J.; Truhlar, D. G., Acc. Chem. Res. 2008, 41 (6), 760-768.

81. Tomasi, J.; Mennucci, B.; Cammi, R., Chem. Rev. 2005, 105 (8), 2999-3093.

82. Tomasi, J.; Persico, M., Chem. Rev. 1994, 94 (7), 2027-2094.

83. Miertus, S.; Scrocco, E.; Tomasi, J., Chem. Phys. 1981, 55 (1), 117-129.

84. Miertus, S.; Tomasi, J., Chem. Phys. 1982, 65 (2), 239-245.

85. Frisch, M. J. T., G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;

Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.;

Page 56: Relativistic Quantum Chemistry Applied to Actinides

34

Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;

Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.;

Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;

Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador,

P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.;

Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul,

A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;

Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,

A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; and

Pople, J. A. Gaussian 03, Revision C.02. 2004.

86. Cocalia, V. A.; Gutowski, K. E.; Rogers, R. D., Coordin. Chem. Rev. 2006, 250 (7-8),

755-764.

87. Gutowski, K. E.; Dixon, D. A., J. Phys. Chem. A 2006, 110 (28), 8840-8856.

88. Moskaleva, L. V.; Kruger, S.; Sporl, A.; Rosch, N., Inorg. Chem. 2004, 43 (13), 4080-

4090.

Page 57: Relativistic Quantum Chemistry Applied to Actinides

35

Preface to Chapter 2

This chapter is based on a manuscript published in the journal “Journal of Physical Chemistry

A”. The full citation of the paper is as follows:

Samuel O. Odoh and Georg Schreckenbach, “Performance of relativistic effective core

potentials in DFT calculations on actinide compounds” Journal of Physical Chemistry A,

2010, 114, 1957.

The actinide atoms are at the tail-end of the periodic table of elements. As such, the use of all-

electron basis sets in theoretical examination of the structure and properties of actinide

complexes are computationally expensive. The use of effective core pseudopotentials (ECPs) in

such calculations reduces the computational burden. In addition, the ECPs are usually

constructed in a careful manner as to include relativistic treatment of the actinide atoms. In this

work, we compared the performance of two types of ECPs in predicting the structure and

reaction free energies of several small actinide complexes against all-electron basis set

calculations.

All the calculations in the published manuscript and compiled in this chapter were carried out by

Samuel O. Odoh. The manuscript was prepared together with Prof. Georg Schreckenbach.

Copyright permissions have been obtained from the American Chemical Society and the other

authors.

Page 58: Relativistic Quantum Chemistry Applied to Actinides

36

Chapter 2: Performance of Relativistic Effective Core Potentials in

DFT Calculations on Actinide Compounds

Abstract

Density functional theory (DFT) calculations using relativistic effective core potentials

(RECPs) have emerged as a robust and fast method of calculating the structural parameters and

energy changes of the thermochemical reactions of actinide complexes. A comparative

investigation of the performance of the Stuttgart small-core and large-core RECPs in DFT

calculations has been carried out. The vibrational frequencies and reaction enthalpy changes of

several uranium (VI) compounds computed using these RECPs were compared to those obtained

using DFT and a four-component one-electron scalar relativistic approximation of the full Dirac

equation with large all-electron basis sets (AE). The relativistic AE method is a full solution of

the Dirac equation with all spin-components separated out. This method gives the „correct‟

answer (with respect to scalar relativity) which should be closest to experimental values when an

adequate density functional is used and in the absence of significant spin-orbit effects. The

small-core RECP generally show better agreement with the four-component scalar- relativistic

AE method than the large-core RECP. We conclude that the 5s, 5p and 5d orbitals are important

in determining the chemistry of actinide complexes. Instances in which large-core RECPs give

better agreement with experimental data are attributed to either experimental uncertainties or

error cancellations.

Page 59: Relativistic Quantum Chemistry Applied to Actinides

37

Introduction

There has been recent progress in the use of quantum mechanical methods to compute the

structural parameters of actinide compounds.1-3

Several approaches have been used to include

relativistic effects, electron correlation and solvation effects into calculations on actinide

complexes.4 To incorporate relativistic effects, various approximations to the full Dirac

Hamiltonian have been used in combination with all-electron (AE) basis sets.5 The literature

contains many examples in which the Douglas-Kroll-Hess (DKH) approximation6-8

, the Zeroth

Order Regular Approximation9-11

(ZORA) and other such approximations have been used to

include relativistic effects in calculations on actinides.12-16

The use of specially-designed

effective core potentials in place of AE basis sets is another means of including relativistic

effects.17

These relativistic effective core potentials (RECPs) and their associated valence basis

sets have been used in calculations at the density functional theory (DFT),16,18

second order

Møller-Plesset Perturbation theory19-21

(MP2) and coupled cluster (CCSD(T))20-22

levels of

theory. There are two popular flavors of RECPs for the uranium atom and other early actinides,

large-core RECPs with the 5f to 7s orbitals treated as valence orbitals (78 core electrons on

uranium) and small-core RECPs which include the 5s, 5p and 5d orbitals in the valence shell as

well (60 core electrons on uranium).

The use of actinide RECPs and associated basis sets in DFT calculations have emerged as

a fast and relatively accurate method of computing the properties of actinide complexes.16,18

The

effects of the density functional(s) of choice on computational results have been examined.21

The

performance of the two types of RECPs mentioned above in DFT computations of the

thermochemical reaction energy changes, vibrational frequencies and bond lengths of actinide

compounds have been compared by various authors. 16,18,23-27

It has been generally noted that

Page 60: Relativistic Quantum Chemistry Applied to Actinides

38

DFT calculations with large-core RECPs often yield less accurate results in comparison to those

using AE methods and also to those using small-core RECPs.

With the release of another set of RECPs by the Stuttgart-Dresden-Bonn (Stuttgart) group

28, Marsden and coworkers compared the performance of the small-core and large-core RECPs

produced specifically by this group.29-30

The vibrational frequencies of various U(VI) compounds

(NUN, NUO+

, CUO, UO3 and UF6) and the enthalpy changes of various reactions involving

several uranium fluorides and oxyfluorides were computed using DFT and the Stuttgart RECPs.

The computed values were statistically compared to experimental values and to results obtained

with the (large-core) Los Alamos RECP and its double-zeta valence basis set (LANL2DZ).31

They concluded from their calculations that similar degrees of accuracy were obtained when

using large-core and small-core RECPs to compute vibrational frequencies. The large-core

RECP was found to give somewhat better enthalpy changes than its small-core counterpart for

the thermochemical reactions that were studied.

This work is a further assessment of the performance of the RECPs from the Stuttgart

group. We have performed DFT calculations using the one-electron four-component scalar

relativistic method32

implemented in the Priroda code33

with large all-electron (AE) basis sets.

The reaction enthalpy changes and vibrational wavenumbers obtained with the AE four-

component method are then compared with those obtained with the RECPs from the Stuttgart

group. In this way, we can separate effects of the relativistic approximation (such as the

comparison of LC- and SC-RECPs) from effects of the model chemistry (DFT exchange-

correlation functional, basis set convergence). Thus fortuitous error cancellations leading to

occasions and possibly conclusions where the large-core RECP performs as well or even better

than the small-core RECP can be identified.

Page 61: Relativistic Quantum Chemistry Applied to Actinides

39

Computational Details

DFT calculations were performed to obtain the optimized geometries and vibrational

frequencies of three linear triatomic molecules (NUN, NUO+ and CUO), UO3 and UF6. The

enthalpy changes associated with the gas phase reactions of several U(VI) compounds were also

computed (reactions 1-5). Three generalized gradient approximation (GGA), PBE34-35

, BPBE35-36

and BLYP36-37

and two hybrid functionals, B3LYP38-39

and PBE034-35

were used in combination

with tight energy and geometry convergence criteria. The AE calculations were carried out with

the Priroda code.33,40

Priroda employs a four-component one-electron scalar relativistic

approximation to the full Dirac equation with all spin-orbit terms separated out and neglected.32

We have previously examined the performance of DFT calculations with the four-component

method implemented in the Priroda program.18,41

Two types of AE basis sets implemented in the

Priroda code were used for all the atoms. These basis sets named L1 and L2 are of double- and

triple-zeta quality respectively for the large component (cc-pVDZ and cc-PVTZ) respectively

40,42. They also include the appropriate kinetically balanced basis sets for the small component. In

Priroda, vibrational constants using the hybrid functionals were obtained with numerical

frequency analysis, whereas analytical frequencies were used for the GGA functionals.

The Stuttgart large-core and small-core RECPs for uranium28, 43-44

and their associated

valence basis sets were used in DFT calculations in both the Gaussian 0345

and NWchem46-47

suites of programs. All g-type functions on the uranium atom were removed from the valence

basis sets. To provide some insight into basis set dependency, the 6-31+G*, 6-311+G*, 6-

311++G(3df,3pd) and aug-cc-pVTZ basis sets were used to describe the hydrogen, nitrogen,

oxygen and fluorine atoms in separate calculations. The same set of density functionals as

employed in Priroda were employed in these calculations. Ultra-fine and xfine grids were used in

Page 62: Relativistic Quantum Chemistry Applied to Actinides

40

all calculations performed using Gaussian 03 and NWChem respectively. Tight self-consistent

field (SCF) and geometry optimization criteria were used in all calculations carried out in

Gaussian 03 and NWChem. Similar settings were employed for both the RECP and AE

calculations. The reaction enthalpy changes at 298K were calculated as the sum of the changes in

the electronic energy and the calculated enthalpy corrections.

Results and Discussion

Vibrational Constants of Triatomic Compounds. NUN, NUO+ and CUO are U(VI)

compounds iso-electronic with UO22+

.The IR spectra of these compounds trapped in argon and

neon matrices have been measured and reported.48-50

Theoretical calculations of the bond-

lengths and vibrational constants of these compounds have been carried out by Pyykkö and co-

workers using large-core quasi-relativistic pseudopotentials at the Hartree-Fock level.51

Gagliardi

and Roos carried out complete active space (CASPT2) calculations using the Stuttgart small-core

RECP on these compounds and obtained results in agreement with experimental data52

. Marsden

and co-workers have also reported the harmonic and an-harmonic vibrational frequencies of

these compounds calculated with DFT and RECPs.29-30

The symmetric and asymmetric

stretching vibrational wavenumbers for NUN, NUO+ and CUO computed using the small-core

and large-core RECPs with the aug-cc-pVTZ ligand basis set are compared to those obtained

using the AE method in Tables 2.1-2.3. Deviations of the values computed using the large- and

small- core RECPs from the AE method are presented graphically in Figure 2.1.

For NUN, vibrational frequencies computed using the small-core RECP show good

agreement with those obtained using the AE method regardless of the basis set or functional

used. The average deviation between the small-core RECP and AE results is 2 cm-1

which is less

Page 63: Relativistic Quantum Chemistry Applied to Actinides

41

than the experimental uncertainty of about 10 cm-1

. However, the values obtained with the large-

core RECP deviate from both the small-core RECP and AE results by 3-25 cm-1

. The observed

deviation is particularly pronounced for the IR-active and more intense anti-symmetric stretching

vibration. The anti-symmetric stretching vibrational frequencies calculated with the PBE and

BPBE GGA functionals and the small-core RECP or AE methods agree very well with the

experimental values, as well as with the high-level CASPT2 ab initio results.52

A similar level of

agreement is maintained even after the inclusion of estimated matrix effects. There is a

systematic 20-25 cm-1

underestimation of the calculated frequencies obtained using the BLYP

functional when compared to the other GGA functionals. This underestimation is observed for all

three methods with which relativistic effects are included. As previously noted, the hybrid

functionals generally over-bind the selected triatomic molecules leading to higher vibrational

frequencies.29-30,51

The general argument that the small-core RECP better reflects the results obtained with

the AE method still remains valid for the calculated frequencies of CUO and NUO+, Figure 2.1.

A close examination of Tables 2.1-2.3 reveals the possibility of significant error cancellation

when comparing the values obtained with the large-core RECP to the experimental values. The

seemingly better agreement between the large-core RECP results obtained using hybrid

functionals and the experimental values is an artifact of the systematic underestimation of the

results obtained with the small-core RECP and AE methods. The comparison of the vibrational

frequencies of NUN, NUO+ and CUO obtained using the large-core and small-core RECPs to

experimental values by the Marsden group is dominated by hybrid functionals.30

This is despite

the fact that the GGA functionals (BP86 and BLYP) that were used gave better agreement with

experimental values when used with the small-core RECP. A bias of their comparison table

Page 64: Relativistic Quantum Chemistry Applied to Actinides

42

towards hybrid functionals for which the values obtained using the large-core RECP

systematically underestimates the small-core RECP values while excluding GGA functionals

Table 2.1: Calculated vibrational frequencies, cm-1

, of NUN.

Experimenta 1077 (1089)

b ω(anti.sym.)

LC RECP SC RECP AE LC RECPc SC RECP

c CASPT2/SC RECP

d

PBE

1021.9

1056.8

1030.0

1081.4

1033.6

1080.5

1015

1072

BPBE

1015.3

1052.8

1025.7

1078.6

1027.3

1075.2

BLYP

987.2

1026.8

997.3

1051.1

1000.8

1050.3

1038

1064

B3LYP

1076.4

1097.5

1072.0

1115.2

1074.4

1113.0

1105

1125

PBE0

1117.8

1134.2

1112.4

1151.2

1115.8

1149.4

1141

1161

a References

48.

b Antisymmetric stretch; experimental vibrational frequencies corrected for

matrix effects are in parenthesis. c Reference

30.

d Reference

52. For Tables 2.1, 2.2 and 2.3, the

calculated frequencies obtained with each functional are given for symmetric (top) and

asymmetric stretching modes.

Page 65: Relativistic Quantum Chemistry Applied to Actinides

43

Table 2.2: Calculated vibrational frequencies, cm-1

, of NUO+.

Experimenta 970 (979) ω(OU)

1119 (1134) ω(NU)

LC RECP SC RECP AE LC RECP c SC RECP

c CASPT2, SC RECP

d

PBE

954.5

1101.5

962.4

1134.8

962.4

1134.8

1023

1134

BPBE

952.0

1097.2

961.0

1131.3

958.7

1124.1

BLYP

920.7

1068.2

931.4

1099.4

932.8

1095.7

932

1078

933

1099

B3LYP

997.3

1167.8

998.6

1186.2

997.2

1180.8

1003

1174

999

1184

PBE0

1038.0

1210.6

1035.0

1230.2

1034.1

1225.0

1042

1218

1038

1229

a Reference

49.

b Experimental vibrational frequencies corrected for matrix effects are in

parenthesis. c Reference

30.

d Reference

52.

(like the popular PBE and BPBE) that give greater agreements with experimental values creates

a situation in which it appears that the large-core RECP performs better or at least as well as the

small-core RECP. Tables 2.1-2.3 contain direct comparisons of the vibrational wavenumbers for

these triatomic compounds computed using the RECPs to the values obtained with a full AE

scalar-relativistic one-electron Dirac equation solution. A cursory look reveals that the RECP

with 60 core electrons always agrees better with the AE method, Figure 2.1. Occasions in which

Page 66: Relativistic Quantum Chemistry Applied to Actinides

44

the large-core RECP appears to give better agreement with the experiment can then be easily

seen as error cancellation due to the hybrid density functionals (typically overestimating the

vibrational wavenumbers) and the nature of the large-core RECP (typically leading to smaller

vibrational wavenumbers than those obtained using the small-core RECP and AE method).

Figure 2.1: Absolute deviations of the vibrational wavenumbers, cm-1

for NUN, NUO+ and

CUO computed using Stuttgart RECPs from the values obtained with the AE four-component

method (small-core RECP in red circles and large-core RECP in black squares. Ligand basis sets

are 6-31+G(d), 6-311+G(d), 6-311++G(3df, 3pd) and aug-cc-pVTZ from Left to Right).

Page 67: Relativistic Quantum Chemistry Applied to Actinides

45

Table 2.3: Calculated vibrational frequencies, cm-1

, of CUO.

Experimenta 872 (881) ω(OU)

1047 (1062)b ω(CU)

LC RECP SC RECP AE LC RECPc SC RECP

c CASPT2,

SC RECPd

PBE

842.4

1076.6

867.0

1094.6

869.9

1086.5

870

1077

BPBE

840.8

1073.4

861.4

1088.0

865.3

1082.0

BLYP

821.3

1051.5

839.7

1055.5

846.1

1050.8

819

1012

842

1064

B3LYP

871.3

1114.1

884.6

1134.4

890.6

1130.6

876

1107

886

1142

PBE0

908.6

1051.5

919.0

1181.6

920.3

1171.9

913

1147

917

1183

a Reference

50.

b Experimental vibrational frequencies corrected for matrix effects are in

parenthesis. c Reference

30.

d Reference

52.

Vibrational Constants of UO3 and UF6. A distorted planar T-shaped structure was obtained for

UO3 in all our calculations, Table 2.4 and Figure 2.2. The calculated 2b2 and 2a1 vibrations of

UO3 and all the vibrational frequencies of the UF6 molecule are presented in Tables 2.5 and 2.6

respectively. The hybrid functionals, B3LYP (and PBE0 for UF6) show better agreement with

experimental results than the GGA functionals used. This is in contrast to the situation observed

Page 68: Relativistic Quantum Chemistry Applied to Actinides

46

for the triatomic molecules above. The performance of the hybrid functionals in calculating the

vibrational frequencies of UO3, UF6, UO2F2 and UO2(OH)2 become relevant when comparing the

reaction enthalpy changes calculated using hybrid and GGA functionals in the section below. A

quick look at Table 2.5 shows that for the functional (B3LYP) that yields the best agreement

with the experimental values, the calculations using the small-core RECP and the AE method

outperform the calculations employing the large-core RECP. Generally, our computed

vibrational wavenumbers are less than those obtained in the work of Marsden when basis sets of

similar triple-ζ quality are used.30

This may be due to different choices of optimization criteria

(we use tight optimization and energy convergence criteria in all our calculations).

Figure 2.2: Distorted Planar T-Shaped Structure of UO3 (g)

The small-core RECP gives structural properties in better agreement with the AE method

than the large-core RECP. As an example, the wavenumbers of the 2a1 and 2b2 vibrations of UO3

computed using the large-core RECP underestimate and overestimate the values obtained using

the two other methods respectively. This relationship between the values computed using the

Page 69: Relativistic Quantum Chemistry Applied to Actinides

47

large-core and small-core RECPs is already obvious from the calculated bond-lengths and bond

angles of the UO3 molecule as shown in Table 2.4. Although, the distorted planar T-shaped

structure was obtained for UO3 in all our calculations, the calculations with the large-core RECP

generally gave longer U-O3 bonds and smaller O1-U-O3 bond angles when compared with the

small-core RECP and AE calculations. In addition, large-core RECP calculations result in shorter

U-O1/O2 bonds and larger O1-U-O2 angles. Thus the smaller and larger values respectively of the

2a1 (U=O3) and asymmetric 2b2 (O1=U=O2) vibrational wavenumbers obtained with the large-

core RECP are not unexpected. Also, the experimental splitting (106 cm-1

) of the 2b2 and 2a1

vibrational modes is better reproduced by the small-core RECP and AE methods (110.14 and

102.48 cm-1

respectively) than by the large-core RECP (139.08 cm-1

) when used with the hybrid

functionals.

The vibrational frequencies of UF6 computed using the small-core RECP are in good

agreement with the previous work of Hirao using a similar RECP and basis sets.53

It can be seen

in Table 2.6 that the large-core RECP gives an erroneous ordering or description of the t1u and

t2g vibrations for all basis sets and density functionals used. This was also observed by Hirao and

co-workers and also in the calculations of Hay and Martin as well as Schreckenbach and co-

workers.1,17

However, the correct wavenumber ordering of the t1u and t2g vibrations was obtained

from calculations using the ZORA Hamiltonian with the BP86 functional and a triple-zeta basis

set performed by Kovács and Konings.54

The calculated vibrational wavenumbers of the UF6

molecule previously obtained by other groups are in agreement with the values shown in Table

2.6.1,17,53,55

Surprisingly however, the erroneous ordering of the t1u and t2g vibrational

wavenumbers was not observed in the large-core RECP calculations of the Marsden group.30

Page 70: Relativistic Quantum Chemistry Applied to Actinides

48

Table 2.4: Calculateda bond lengths (in Å) and bond angles of gaseous UO3.

LC RECP SC RECP AE

PBE 1.798 1.863

100.4 159.3

1.820 1.858

101.7 156.7

1.823 1.859

102.4 155.2

BPBE 1.799 1.864

100.5 159.0

1.821 1.858

101.9 156.2

1.827 1.862

102.9 154.2

BLYP 1.817 1.881

101.8 156.5

1.844 1.876

104.3 151.3

1.847 1.877

105.0 150.1

B3LYP 1.780 1.852

99.8 160.4

1.803 1.848

100.5 159.0

1.807 1.850

101.1 157.8

PBE0 1.761 1.834

98.5 163.1

1.780 1.830

98.9 162.3

1.783 1.832

99.4 161.3

a The parameters are ordered as r(U-O1/O2), r(U-O3), <O1-U-O3 and <O1-U-O2 respectively

Page 71: Relativistic Quantum Chemistry Applied to Actinides

49

Table 2.5: Calculated vibrational frequencies, cm-1

, of UO3.

Experimenta 760.3 (768) 2a1 (U=O3)

865.3 (875)b

2b2 (O1=U=O2)

LC RECP SC RECP AE LC RECPc SC RECP

c ADF/ZORA

d

PBE

727.1

854.0

751.5

843.6

757.7

843.6

758.3

849.0

BPBE

726.7

853.9

751.6

841.7

755.0

838.8

BLYP

700.5

823.3

729.4

806.7

736.7

809.2

712

840

727

809

B3LYP

756.2

891.0

775.3

879.3

781.6

879.9

766

906

779

891

PBE0

788.8

927.6

803.9

919.5

809.7

918.0

798

942

812

935

a Experimental vibrational wavenumbers corrected for matrix effects are in parenthesis.

b

Reference 56

. c Reference

30.

d Reference

21.

Several general trends observed in the calculated wavenumbers of the triatomic

molecules and previously by other workers were also observed in the computed values of the

vibrational frequencies of UO3, UF6, UO2F2, and UO2(OH)2. The GGA functionals under-bind

the atoms in these molecules leading to vibrational frequencies lower than those obtained with

hybrid functionals. Also, of the GGA functionals considered, the BLYP functional systematically

underestimates the other GGAs (PBE and BPBE) by approximately 10-30 cm-1

. This is most

Page 72: Relativistic Quantum Chemistry Applied to Actinides

50

noticeable for stretching vibrations. The small-core RECP is always in better agreement with the

four-component scalar-relativistic AE method employed in this work. The deviations of the

vibrational wavenumbers of UO3 and UF6 computed with the large-core and small-core RECPs

from those obtained using the AE method are presented in Figure 2.3.

Figure 2.3: Absolute deviations of the vibrational wavenumbers, cm-1

, for UO3 and UF6

computed using Stuttgart RECPs from the values obtained with the AE four-component method

(small-core RECP in red circles and large-core RECP in black squares. Ligand basis sets are 6-

31+G(d), 6-311+G(d), 6-311++G (3df, 3pd) and aug-cc-pVTZ from left to right).

Page 73: Relativistic Quantum Chemistry Applied to Actinides

51

Table 2.6: Calculated vibrational frequencies, cm-1

, of UF6.

t2u t1u t2g eg t1u a1g

Experimenta 143 186 200 540 634 672

Large-core RECP

B3LYP 143.5 187.9 184.2 515.2 613.4 645.8

PBE0 148.5 189.8 188.5 525.6 631.3 669.4

PBE 139.5 179.7 173.5 501.1 591.4 609.1

BPBE 140.3 180.6 173.5 500.4 590.7 608.2

BLYP 136.8 178.5 170.2 490.4 576.0 591.7

Small-core RECP

B3LYP 137.4 181.6 191.9 527.3 610.3 654.5

PBE0 138.7 182.2 193.0 537.0 627.6 676.4

PBE 130.0 173.1 182.6 516.4 590.8 624.4

BPBE 130.1 173.6 183.3 516.4 590.3 623.7

BLYP 129.6 172.3 183.2 506.7 575.8 607.9

AE-4-Component

B3LYP 137.7 183.0 198.1 536.5 618.1 662.7

PBE0 139.1 184.5 199.8 545.5 635.8 683.6

PBE 130.3 175.6 187.4 525.4 599.1 632.1

BPBE 131.9 176.8 189.4 520.9 593.5 626.6

BLYP 130.0 175.4 188.6 517.0 585.7 617.2

a Reference

57

Page 74: Relativistic Quantum Chemistry Applied to Actinides

52

Generally, the vibrational wavenumbers computed using the small-core RECP tend to

converge toward the values computed with the AE method as the quality of the basis set is

increased. The structural parameters computed with the small-core RECP with a triple-zeta basis

set can be used in lieu of an AE calculation with some degree of confidence, Figures 2.1 and

2.3, Tables 2.1-2.6.

Enthalpy of Reaction. The enthalpy changes associated with the thermochemical reactions

studied in this work have been reported experimentally by Privalov and co-workers as -311

kJ/mol, 65 kJ/mol, 187 kJ/mol, -184 kJ/mol and 435 kJ/mol for reactions 1-5 respectively.19

2UO3(g) + UF6(g) → 3UO2F2(g) 1

UO2F2(g) + 2H2O(g) → UO2(OH) 2(g) + 2HF(g) 2

UF6(g) + 2H2O(g) → UO2F2(g) + 4HF(g) 3

UO3(g) + H2O(g) → UO2(OH)2(g) 4

UF6(g) + 3H2O(g) → UO3(g) + 6HF(g) 5

It should be noted that there are uncertainties of 15-50 kJ/mol associated with the experimental

data obtained. There have been several attempts to calculate the enthalpy changes associated

with these reactions. The Privalov group also calculated the reaction enthalpy changes using

MP2, B3LYP and CCSD(T) with the Stuttgart small-core RECP. The inclusion of diffuse

functions in the basis sets was found to result in more accurate enthalpy changes by the Marsden

group30

. Shamov and co-workers computed the enthalpy changes of reactions 1, 4 and 5, among

others using the same AE method used in this work21

. They compared the AE reaction enthalpy

changes to those obtained using the small-core RECP and ZORA/ADF methods. Actually, a

Page 75: Relativistic Quantum Chemistry Applied to Actinides

53

close examination of Table 2.7 in their work reveals great agreement between the results

obtained with the AE method, ZORA/ADF method and the small-core RECP when larger basis

sets are used.

The calculated enthalpy changes obtained using the AE method with the triple-zeta

quality L2 basis and the Stuttgart RECPs with the aug-cc-pVTZ basis set used on oxygen,

fluorine and hydrogen are presented in Tables 2.7. From Figure 2.4, it is obvious that the

enthalpy changes calculated using the small-core RECP are in much better agreement with the

AE method than those obtained using the large-core RECP. The small-core and large-core

RECP enthalpy changes are generally within 0-50 and 10-100 kJ/mol of the values obtained

using the AE method, respectively. The enthalpy changes computed with the small-core RECP

and AE method converge dramatically towards the experimental data as the number of diffuse

functions in the basis set is increased. This is also observed for calculations in which the large-

core RECP was used, albeit not uniformly. This is in line with previous observation of

improvements in the calculated enthalpy change when diffuse functions are added to the basis

set.

The reaction enthalpy changes calculated using hybrid functionals with the small-core

RECP and AE method for reactions 1, 2 and 4 are in good agreement with the experimental data.

Privalov and co-workers suggested that the rather large errors in the calculated values obtained

for reactions 2, 3 and 5 arise from the poor description of HF in these reactions. This was refined

by Marsden and co-workers as possibly due to the performance of the functionals in calculating

Page 76: Relativistic Quantum Chemistry Applied to Actinides

54

Table 2.7: Calculated enthalpy changes*, kJ/mol for Reactions 1-5 computed using Stuttgart

RECPs and the AE four-component method.

Reaction Method Functional

B3LYP PBE0 PBE BPBE BLYP

1 LC RECP -326.5 -355.8 -218.8 -218.6 -210.5

SC RECP -281.4 -305.9 -197.4 -198.8 -191.5

AE-Method -275.9 -302.0 -196.8 -198.4 -187.7

Experiment -311

2 LC RECP 106.7 103.0 98.9 100.0 102.5

SC RECP 95.9 91.2 89.5 90.9 95.0

AE-Method 95.2 90.6 89.8 92.5 94.4

Experiment 65

3 LC RECP 181.4 166.6 220.7 197.2 218.4

SC RECP 248.4 227.7 282.7 261.8 286.8

AE-Method 274.8 248.4 312.6 279.7 321.4

Experiment 187

4 LC RECP -147.2 -158.2 -120.7 -108.0 -112.0

SC RECP -169.0 -175.6 -151.1 -139.4 -144.2

AE-Method -180.1 -184.6 -164.9 -146.5 -160.1

Experiment -184

5 LC RECP 435.5 427.7 440.3 405.2 432.9

SC RECP 513.3 494.5 524.3 492.1 526.0

Page 77: Relativistic Quantum Chemistry Applied to Actinides

55

AE-Method 550.1 523.7 567.4 518.8 575.9

Experiment 435

* Experimental values are from reference

19. The aug-cc-pVTZ ligand basis was used in the

RECP calculations while the triple-zeta L2 basis was used in the AE calculations.

Figure 2.4: Absolute deviations of the enthalpy changes, kJ/mol for reactions 1-5 computed

using Stuttgart RECPs from the values obtained with the AE four-component method (small-

core RECP in red circles and large-core RECP in Black Squares. Ligand basis sets are 6-

31+G(d), 6-311+G(d), 6-311++G (3df, 3pd) and aug-cc-pVTZ from left to right)

Page 78: Relativistic Quantum Chemistry Applied to Actinides

56

the electron affinity of the F atom.30

It is worthy of note that the calculations that employ the

small-core RECP and AE method perform very well for the reaction with no fluoride involved

(Reaction 4). It could be argued that treatment of the fluorides is a pathologic problem for these

methods. The large discrepancies between the experimental values and those obtained using

hybrid functionals with small-core RECP and AE method is an indication of the poor

performance of approximate DFT in predicting the enthalpy changes of Reactions 2, 3, and 5. It

has been previously shown that CCSD(T) calculations using either a four-component scalar-

relativistic method with all-electron basis sets or the small-core RECP give results in greater

agreement with experimental values.21,30

Although, the energy changes obtained in calculations

using the large-core RECP give exceptional agreement with experimental values for Reactions 3

and 5, it is obvious that this is due to fortuitous error cancellation. In any case, the right

comparison is between the values obtained with the AE and RECP methods and not between the

results of the RECP methods and the experimental values.

Conclusions

We have carried out a comparative study of the performance of the Stuttgart small-core

and large-core RECPs against an all-electron method that employs DFT and a four-component

one-electron scalar-relativistic approximation. The vibrational wavenumbers of several uranium

compounds and the enthalpy changes associated with the thermochemical reactions of several

uranium fluorides and oxyfluorides were calculated using the RECPs and the AE method.

The structural parameters and reaction enthalpy changes computed using DFT and the

four-component one-electron scalar relativistic AE method are the „correct DFT‟ values. Granted

these values may deviate significantly from experimental data as has been previously shown, the

Page 79: Relativistic Quantum Chemistry Applied to Actinides

57

values computed using the small-core RECP are generally in greater agreement than those

obtained using the large core RECP. The importance of the 5s, 5p and 5d orbitals in describing

the chemistry of the selected actinide compounds have been made obvious. As has been

previously discussed this somewhat counterintuitive result can be understood as follows. In ECP

methods, pseudo-orbitals are used for the valence orbitals that, by construction, lack the nodal

structure (core wiggles) arising from the orthogonality requirements with respect to the core

shells.4 By adding the 5s, 5p and 5d orbitals to the calculations, we effectively reintroduce the

outermost such core wiggle to the higher orbitals. This core wiggle may well stretch into regions

of space that are relevant to bonding, and it would thus be required for an accurate description of

the chemistry of these systems. We should note that the neglect of core-polarization and

correlation has been put forward as an alternative explanation of the differences between large-

and small-core ECPs.28

Discrepancies between the experimental data and the values computed with the AE could

be due to non-suitability of approximate DFT in studying the selected system, significant spin-

orbit effects or experimental errors. Instances in which the large-core RECP appear to give better

results than the small-core RECP and AE method are in our opinion due to error cancellations

(basis sets, functionals and RECPs errors). One could, of course, argue that accurate results

obtained through error cancellations are still valuable results. While this argument has its merits,

there is no guarantee, as far as we can tell, that this error cancellation is transferable between

different types of uranium compounds or to different oxidation states of the uranium atom. Thus

we conclude based on the findings in this work that the use of large-core RECPs for calculations

on actinide complexes is not recommended anymore.

Page 80: Relativistic Quantum Chemistry Applied to Actinides

58

References

1. Schreckenbach, G.; Hay, P. J.; Martin, R. L., J. Comput. Chem. 1999, 20, 70.

2. Vallet, V.; Macak, P.; Wahlgren, U.; Grenthe, I., Theor. Chem. Acc. 2006, 115, 145.

3. Kaltsoyannis, N.; Hay, P. J.; Li, J.; Blaudeau, J. P.; Bursten, B. E., Theoretical Studies of

the Electronic Structure of Compounds of the Actinides. In The Chemistry of the Actinide and

Transactinide Elements, 3rd ed, 3 ed.; Morss, L. R.; Edelstein, N. M.; Fuger, J.; Katz, J. J., Eds.

Springer: Dordrecht, The Netherlands, 2006; Vol. 3, pp 1893.

4. Schreckenbach, G.; Shamov, G. A., Acc. Chem. Res. 2009, in print 10.1021/ar800271r.

5. Dirac, P. A. M., Proc. R. Soc. London, Ser. A 1928, 117, 610.

6. Douglas, M.; Kroll, N. M., Ann. Phys. 1974, 82, 89.

7. Hess, B. A., Phys. Rev. A: At. Mol. Opt. Phys. 1986, 33, 3742.

8. Reiher, M., Theor. Chem. Acc. 2006, 116, 241.

9. Faas, S.; Snijders, J. G.; van Lenthe, J. H.; van Lenthe, E.; Baerends, E. J., Chem. Phys.

Lett. 1995, 246, 632.

10. van Lenthe, E., J. Comput. Chem. 1999, 20, 51.

11. van Lenthe, E.; Baerends, E. J.; Snijders, J. G., J. Chem. Phys. 1993, 99, 4597.

12. Peralta, J. E.; Batista, E. R.; Scuseria, G. E.; Martin, R. L., J. Chem. Theory Comput.

2005, 1, 612.

13. Cao, X. Y.; Moritz, A.; Dolg, M., Chem. Phys. 2008, 343, 250.

14. van Wüllen, C., J. Comput. Chem. 1999, 20, 51.

15. Infante, I.; Visscher, L., J. Comput. Chem. 2004, 25, 386.

16. Garcia-Hernandez, M.; Lauterbach, C.; Krüger, S.; Matveev, A.; Rösch, N., J. Comput.

Chem. 2002, 23, 834.

Page 81: Relativistic Quantum Chemistry Applied to Actinides

59

17. Hay, P. J.; Martin, R. L., J. Chem. Phys. 1998, 109, 3875.

18. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109, 10961.

19. Privalov, T.; Schimmelpfennig, B.; Wahlgren, U.; Grenthe, I., J. Phys. Chem. A 2002,

106, 11277.

20. Schimmelpfennig, B.; Privalov, T.; Wahlgren, U.; Grenthe, I., J. Phys. Chem. A 2003,

107, 9705.

21. Shamov, G. A.; Schreckenbach, G.; Vo, T. N., Chem. Eur. J. 2007, 13, 4932.

22. Rotzinger, F. P., Chem. Eur. J. 2007, 13, 800.

23. Straka, M.; Kaupp, M., Chem. Phys. 2005, 311, 45.

24. de Jong, W. A.; Harrison, R. J.; Nichols, J. A.; Dixon, D. A., Theor. Chem. Acc. 2001,

107, 22.

25. Batista, E. R.; Martin, R. L.; Hay, P. J., J. Chem. Phys. 2004, 121, 11104.

26. Vetere, V.; Maldivi, P.; Adamo, C., J. Comput. Chem. 2003, 24, 850.

27. Tsushima, S.; Uchida, Y.; Reich, T., Chem. Phys. Lett. 2002, 357, 73.

28. Lim, I. S.; Stoll, H.; Schwerdtfeger, P., J. Chem. Phys. 2006, 124, 034107.

29. Clavaguera-Sarrio, C.; Ismail, N.; Marsden, C. J.; Begue, D.; Pouchan, C., Chem. Phys.

2004, 302, 1.

30. Iche-Tarrat, N.; Marsden, C. J., J. Phys. Chem. A 2008, 112, 7632.

31. Hay, P. J.; Wadt, W. R., J. Chem. Phys. 1985, 82, 270.

32. Dyall, K. G., J. Chem. Phys. 1994, 100, 2118.

33. Laikov, D. N.; Ustynyuk, Y. A., Russ. Chem. Bull. 2005, 54, 820.

34. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1996, 77, 3865.

35. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1997, 78, 1396.

Page 82: Relativistic Quantum Chemistry Applied to Actinides

60

36. Becke, A. D., Phys. Rev. A: At. Mol. Opt. Phys. 1988, 38, 3098.

37. Lee, C. T.; Yang, W. T.; Parr, R. G., Phys. Rev. B: Condens. Matter 1988, 37, 785.

38. Becke, A. D., J. Chem. Phys. 1993, 98, 5648.

39. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J., J. Phys. Chem. 1994, 98,

11623.

40. Laikov, D. N. Priroda Code 6. 2006.

41. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2006, 110, 9486.

42. Laikov, D. N., Chem. Phys. Lett. 2005, 416, 116.

43. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74, 1245.

44. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100, 7535.

45. Frisch, M. J. T., G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;

Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.;

Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;

Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.;

Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;

Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador,

P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.;

Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul,

A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;

Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,

A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; and

Pople, J. A. Gaussian 03, Revision C.02. 2004.

Page 83: Relativistic Quantum Chemistry Applied to Actinides

61

46. Bylaska, E. J. d. J., W. A.; Govind N.; Kowalski K.; Straatsma, T. P.; Valiev, M.; Wang,

D.; Apra, E.; Windus, T. L.; Hammond, J.; Nichols, P.; Hirata, S.; Hackler, M. T.; Zhao, Y.; Fan,

P. -D.; Harrison, R. J.; Dupuis, M.; Smith, D. M. A.; Nieplocha, J.; Tipparaju, V.; Krishnan, M.;

Wu, Q.; Van Voorhis, T.; Auer, A. A.; Nooijen, M.; Brown, E.; Cisneros, G.; Fann, G. I.;

Fruchtl, H.; Garza, J.; Hirao, K.; Kendall, R.; Nichols, J. A.; Tsemekhman, K.; Wolinski, K.;

Anchell, J.; Bernholdt, D.; Borowski, P.; Clark, T.; Clerc, D.; Dachsel, H.; Deegan, M.; Dyall,

K.; Elwood, D.; Glendening, E.; Gutowski, M.; Hess, A.; Jaffe, J.; Johnson, B.; Ju, J.;

Kobayashi, R.; Kutteh, R.; Lin, Z.; Littlefield, R.; Long, X.; Meng, B.; Nakajima, T.; Niu, S.;

Pollack, L.; Rosing, M.; Sandrone, G.; Stave, M.; Taylor, H.; Thomas, G.; van Lenthe, J.; Wong,

A.; and Zhang, Z. NWChem, A Computational Chemistry Package for Parallel Computers,

Version 5.1". 2007.

47. Kendall, R. A.; Apra, E.; Bernholdt, D. E.; Bylaska, E. J.; Dupuis, M.; Fann, G. I.;

Harrison, R. J.; Ju, J. L.; Nichols, J. A.; Nieplocha, J.; Straatsma, T. P.; Windus, T. L.; Wong, A.

T., Comput. Phys. Commun. 2000, 128, 260.

48. Hunt, R. D.; Yustein, J. T.; Andrews, L., J. Chem. Phys. 1993, 98, 6070.

49. Zhou, M. F.; Andrews, L., J. Chem. Phys. 1999, 111, 11044.

50. Andrews, L.; Liang, B. Y.; Li, J.; Bursten, B. E., Angew. Chem. Int. Ed. 2000, 39, 4565.

51. Pyykkö, P.; Li, J.; Runeberg, N., J. Phys. Chem. 1994, 98, 4809.

52. Gagliardi, L. A.; Roos, B. O., Chem. Phys. Lett. 2000, 331, 229.

53. Han, Y. K.; Hirao, K., J. Chem. Phys. 2000, 113, 7345.

54. Kovács, A.; Konings, R. J. M., J. Mol. Struc.-Theochem 2004, 684, 35.

55. Han, Y. K., J. Comput. Chem. 2001, 22, 2010.

56. Zhou, M. F.; Andrews, L.; Ismail, N.; Marsden, C., J. Phys. Chem. A 2000, 104, 5495.

Page 84: Relativistic Quantum Chemistry Applied to Actinides

62

57. McDowell, R. S.; Asprey, L. B.; Paine, R. T., J. Chem. Phys. 1974, 61, 3571.

Page 85: Relativistic Quantum Chemistry Applied to Actinides

63

Preface to Chapter 3

This chapter is based on a manuscript published in the journal “Journal of Physical Chemistry

A”. The full citation of the paper is as follows:

Samuel O. Odoh and Georg Schreckenbach, “Theoretical study of the structural

properties of Pu(IV) and Pu(VI) Complexes” Journal of Physical Chemistry A, 2010,

114, 1957.

Density functional theory (DFT) is a single reference approach. The implication of this is that it

is not a fully rigorous approach for studying multi-reference systems, i.e. systems in which the

ground electronic state contains two or more nearly degenerate determinants. Open-shell actinide

ions such as Pu(VI) and Pu(IV) are f2 and f

4 systems.

All the calculations in the published manuscript and compiled in this chapter were carried out by

Samuel O. Odoh. The manuscript was prepared together with Prof. Georg Schreckenbach.

Copyright permissions have been obtained from the American Chemical Society and the other

authors.

Page 86: Relativistic Quantum Chemistry Applied to Actinides

64

Chapter 3: Theoretical Study of the Structural Properties of

Plutonium IV and VI Complexes

Abstract

The structural properties of several plutonium (IV) and (VI) complexes have been

examined in the gaseous and aqueous phases using Kohn-Sham density functional theory

calculations with scalar relativistic effective core potentials and the polarizable continuum

solvation model. The aquo and nitrate complexes of PuO22+

and Pu4+

were considered in addition

to the aquo-chloro complexes of PuO22+

. The nitrate and chloro- complexes formed with

triphenylphosphine oxide (TPPO) and tributylphosphate (TBP) respectively were also studied.

The structural parameters of the plutonyl complexes were compared to their uranyl and neptunyl

analogues. The bond lengths and vibrational frequencies of the plutonyl complexes can generally

be computed with sufficient accuracy with the pure PBE density functional with shorter bond

lengths being predicted by the B3LYP functional. The structural parameters of the [PuO2Cl2L2]

systems formed with TPPO and TBP as well as the aqueous [PuO2Cl2(H2O)3] complex are

matched to previous experimental results. Overall, the inclusion of ligands in the equatorial

region results in significant changes in the stretching frequency of the plutonyl group. The

structural features of the plutonyl (VI) systems are rather similar to those of their 5f0 uranyl and

5f1 neptunyl counterparts. For the Pu(IV) aquo and nitrate complexes, the average of the

calculated Pu-OH2 and Pu-Onitrate bond lengths are generally within 0.04 Å of the reported

experimental values. Overall Kohn-Sham DFT can be used successfully in predicting the

structures of this diverse set of Pu(VI) and Pu(IV) complexes.

Page 87: Relativistic Quantum Chemistry Applied to Actinides

65

Introduction

The speciation and mobility of actinide complexes in the environment depend on the

prevailing pH, ligand concentration as well as the presence of oxidizing agents. It is therefore

important to study the structural, redox and photochemical properties of actinide compounds

using experimental and theoretical approaches. On the theoretical side, there has been

tremendous progress in the use of quantum mechanical methods in probing the structure and

electronic properties of actinide complexes.1-24

The incorporation of relativistic effects, electron

correlation, and when needed, solvation effects have drastically increased the accuracy and

utility of such methods. Increasingly, Kohn-Sham density functional theory (DFT) has emerged

as a method of choice for carrying out calculations on uranium (VI) complexes as a result of its

speed, good treatment of electron correlation and the strongly single-reference character of these

5f0 compounds. The presence of valence shell electrons in 5f

n (n≠0) complexes often results in a

multitude of low-lying states that are nearly degenerate with the electronic ground state. In such

cases, there is the possibility that single reference approaches will predict the wrong electronic

ground state, or would have large spin contamination or poor energy convergence attributes with

the 5f-electrons hopping between the nearly-degenerate orbitals. For a sizeable fraction of

actinide complexes, the wavefunction is distinctly multi-reference in character making single

reference approaches like Kohn-Sham DFT (simply DFT from here onwards) at best less

rigorous or at worst unsuitable. Increasingly however, scalar-relativistic DFT has been shown to

successfully predict the structural properties, thermochemistry and bonding of open-shell

actinide systems.20, 22, 25-27

This has often been attributed to the atomic character of the occupied

5f orbitals ensuring little difference in the general chemical bonding schemes of the different

electronic states formed by the different 5f occupations.27

Page 88: Relativistic Quantum Chemistry Applied to Actinides

66

For the case of the plutonyl (VI) moiety, PuO22+

, DFT predicts the 3Hg ground state in

agreement with more sophisticated correlated ab initio methods.28

This, in addition makes the

accuracy of DFT in predicting the geometrical structures of plutonyl complexes unsurprising.25-

26, 29 It should however be noted that the calculated structural parameters obtained from scalar-

relativistic DFT are accurate even though the true 3Hg ground state is an admixture of equal

proportions of two determinants each with a normalization factor of 1/√2. The calculated

structural properties for plutonyl complexes obtained using DFT can therefore be compared to

available experimental data to further assess the performance of density functionals in predicting

their speciation and equilibrium structures. In their calibration work, Ismail et al.28

obtained

similar Pu=Oyl bond lengths and vibrational frequencies for the PuO22+

group with the B3LYP

functional and the approximate quadratic coupled cluster approach. Indeed, it appears that the

structural features obtained using the B3LYP functional are also in very good agreement with

those obtained with the scalar relativistic and spin-orbit corrected variants of the complete active

space, CASSCF, approach after correcting for dynamic correlations using second-order

perturbation theory, CASPT2.30

A more general work by Clavaguera-Sarrio et al.27

found out

that the description of the electronic ground state, structural and vibrational properties of open-

shell actinide systems are described by DFT with accuracy on par with those obtained with the

CASPT2 approach. The structures of carbonate and aquo complexes of Pu(VI) have also been

examined using DFT in conjunction with perturbation theory and coupled cluster approaches.20,

25-26, 31 These works showed that DFT methods can sufficiently provide a quantitative estimate of

the equilibrium number of equatorial ligands in acidic aqueous conditions.

Plutonium complexes are significantly less studied with experimental techniques than

their uranium counterparts for reasons of toxicity and safety. The possibility of producing

Page 89: Relativistic Quantum Chemistry Applied to Actinides

67

nuclear weapons from stolen plutonium is viewed as a significant risk by many global powers.

Theoretical calculations on plutonium complexes can then be used to either complement

available experimental data or to bridge the gaps in our knowledge of these complexes. On the

experimental side, Gaunt et al.32

have recently characterized the crystal structure of the plutonyl

(VI) dinitrate, [PuO2(NO3)2(H2O)2], complex using X-ray diffraction and spectroscopic methods.

In contrast, they found only negligible formation of the dinitrate in aqueous solutions with the

mononitrates, [PuO2(NO3)(H2O)x]+, being dominant despite the weak nitrate complexation by the

plutonyl entity. There is significantly less theoretical data on plutonyl nitrate complexes in

contrast to their counterpart aquo,20, 25, 29

carbonate,33

hydroxo31

and plutonium (VI) fluoride4, 9

complexes. This is most likely due to the weak complexation of nitrates by actinyl groups in

general despite the importance of nitrates and nitric acid in industrial nuclear processes.

There is however available experimental data on the structures of actinide (IV) nitrate

[An(NO3)n(H2O)m]4-n

, complexes making assessment of the performance of theoretical methods

easier.34-35

The Pu4+

cation has a 5f4 electronic configuration and is known to coordinate with 5

or 6 nitrate ligands with both species existing in aqueous solutions at very high nitrate

concentrations. Recently, Horowitz and Marston used scalar-relativistic calculations to

accurately predict the structures of U(IV), Np(IV) and Pu(IV) aquo complexes.29

In addition,

they obtained relatively accurate redox potentials for the Pu(III)/Pu(IV) redox couple using spin-

orbit corrected DFT. This and other reports20, 25-27, 31

would suggest that DFT should be able to at

the least correctly predict the structural features and to some extent correctly predict the

speciation of Pu(IV) nitrate complexes.

In this work, the reliability of DFT calculations in predicting the geometrical features and

vibrational frequencies of diverse gaseous and aqueous phase plutonium (IV) and (VI)

Page 90: Relativistic Quantum Chemistry Applied to Actinides

68

complexes is further examined. The aim is to extend the application of DFT in predicting the

structures of plutonium complexes beyond the extensively examined aquo, carbonate and

hexahalide complexes. This is done by appraising its capability in studying other plutonium

complexes not extensively studied using theoretical approaches. In this work, the structures of

the pentaaquo, aquo-chloro and nitrate complexes of PuO22+

are examined. The dinitrate and

dichloro complexes formed with organic ligands such as tributylphosphate (TBP) and

triphenylphosphine oxide (TPPO) were also examined in addition to the [AnO2Cl4]2-

and their

cesium salts, Cs2[AnO2Cl4] (An=U, Np and Pu). These diverse ligand environments should allow

us to gain insights into the nature of the Pu-ligand interaction, the vibrational spectra of these

molecules as well as the ability of DFT to provide sufficiently accurate structures for these

molecules. The comparison of the uranium, neptunium and plutonium complexes will also

provide insights into any structural changes down the actinide series. Only the aquo,

[Pu(H2O)8]4+

and nitrate [Pu(NO3)n(H2O)m]4-n

, complexes are considered for the +4 oxidation

state.

Computational Details

The geometries of all the complexes in this work were fully optimized without symmetry

constraints in the gaseous and when needed aqueous phases. The optimized structures were

confirmed as local minima structures using vibrational frequency analysis. The calculated

vibrational frequencies were obtained using the harmonic approximation. Although this work is

mainly focused on plutonium complexes, we have compared some of these complexes with their

uranium and neptunium counterparts. We have only compared compounds in which the actinide

atoms possess the same oxidation states. Scalar relativistic effective core pseudopotentials

(RECPs) were used in this work. The RECP calculations were carried out with the Gaussian 0336

Page 91: Relativistic Quantum Chemistry Applied to Actinides

69

suite of programs. The Stuttgart small-core37-38

(60 core electrons represented by a

pseudopotential) RECPs and associated valence basis sets were used for the actinide atoms while

the aug-cc-pVDZ basis was used on all other atoms. The uranium, neptunium and plutonium

valence basis sets were augmented with diffuse f-type functions having exponent of 0.005, 0.005

and 0.05 respectively. All g-type functions were however removed from the actinide valence

basis sets. This small-core pseudopotential has been shown to result in very good agreements

with all-electron basis set calculations as well as experimental structural parameters and reaction

energies.13, 19-20, 22

To examine the basis set dependence of the energetics of the complexes

studied in this work, the aug-cc-pVTZ basis was also used on the hydrogen, nitrogen, oxygen

and chlorine atoms in the molecular structures optimized with the double-ζ basis. The hybrid

B3LYP39-40

and pure DFT PBE41-42

functionals were used in these calculations. The numerical

integration of the exchange-correlation portion of the density functionals were carried out using

ultra-fine grids. All the complexes studied in this work had negligible spin contaminations in the

unrestricted DFT ground states. Deviations of more than 3.5% from the expected S(S+1) values

were taken as the hallmark of „significant‟ spin contamination. The solution phase calculations

were carried out with the polarizable continuum solvation (PCM) model.43

All solution phase

calculations were carried out in water. The default atomic radii in the Gaussian 03 code were

used in the calculations. It should be mentioned that no imaginary frequencies were obtained for

the complexes optimized in aqueous solution using the PCM model. The gas phase and aqueous

phase frequency analyses calculations were carried out analytically. Tight geometry optimization

and energy convergence criteria were stipulated in all the calculations. The various gas-phase

optimized structures for each molecule were used as starting structures for optimization using the

Page 92: Relativistic Quantum Chemistry Applied to Actinides

70

PCM model. Only the bidentate coordination mode was considered for all the nitrate complexes

studied in this work.34-35

Results and Discussion

Bare PuO22/1+

and aquo plutonyl (VI/V) complexes. The calculated structural parameters of

the bare PuO22/1+

and pentaaquo plutonyl (VI/V) complexes obtained using RECPs are presented

in Table 3.1. Starting from the naked plutonyl dication, PuO22+

, the calculated Pu=O bond length

obtained with the B3LYP functional in the gas phase is in good agreement with those obtained

using the CASPT2 and AQCC approaches.28, 30

Also, there is good agreement between the

calculated stretching vibrational frequencies of the plutonyl entity obtained using the B3LYP

functional and these more sophisticated approaches. The use of the PBE functional results in

longer Pu=O bonds and consequently lower vibrational frequencies. The decrease in the

vibrational frequencies caused by employing a GGA functional is on the order of 78-85 cm-1

.

Similar effects have been observed in the performance of the PW91 and B3LYP functionals in

predicting the structural features of PuN2.27

The Pu=O bond of the bare PuO22+

is slightly longer

by about 0.03 Å when the B3LYP functional and PCM solvation model are employed. This

increase in length results in stretching vibrational frequencies of similar magnitude as those

obtained when the GGA functional is used in the gas-phase. Employing the GGA functional with

the PCM solvation model results in very good agreement between the calculated and

experimental Pu=O bond lengths. This also brings about very good agreement (within 6 cm-1

and

2 cm-1

respectively) between the calculated and experimental symmetric and asymmetric

stretching vibrational frequencies. It seems that employing the GGA functional with the PCM

solvation model in calculations on the bare plutonyl entity, although questionable on theoretical

Page 93: Relativistic Quantum Chemistry Applied to Actinides

71

Table 3.1: The calculated bond lengths (Å) and plutonyl vibrational frequencies (cm-1

) of the

bare PuO22/1+

and actinyl aquo PuO2(H2O)52/1+

systems obtained using DFT in the gaseous and

aqueous phases

System Parameter Gas Solution Expt.a

PBE B3LYP PBE B3LYP

PuO22+

Pu=Oyl 1.704 1.669 1.738 1.703

υsymm 919.3 1006.1 842.0 916.1

υasymm 1044.1 1122.7 964.1 1024.1

PuO2(H2O)52+

Pu=Oyl 1.741 1.712 1.748 1.718 1.74

Pu-OH2 2.465 2.460 2.441 2.435 2.41

υsymm 848.5 918.0 829.7 898.3 835

υasymm 971.0 1031.1 937.5 991.4 962

PuO21+

Pu=Oyl 1.737 1.723 1.788 1.776

υsymm 870.2 911.4 761.9 792.4

υasymm 968.4 1007.6 839.3 863.4

PuO2(H2O)51+

Pu=Oyl 1.788 1.774 1.801 1.788 1.81

Pu-OH2 2.583 2.590 2.565 2.568 2.47

Page 94: Relativistic Quantum Chemistry Applied to Actinides

72

υsymm 782.2 817.2 753.2 786.0 748

υasymm 874.0 908.0 818.0 843.2

a References

61, 64.

grounds, is sufficient in predicting the structural parameters in acidic solutions. The agreement

between the experimental and calculated Pu=O bond lengths and vibrational frequencies suggest

that the effect of explicitly adding aquo ligands in the first hydration sphere around the plutonyl

dication should be minimal. Comparison of the aqueous phase structures obtained for PuO22+

and

[PuO2(H2O)5]2+

using the GGA functional, Table 3.1, shows a minute elongation of the Pu=Oyl

bonds by about 0.01 Å, suggesting that the Pu-OH2 interaction is very weak. There is also a

minimal decrease in the Oyl=Pu=Oyl angle by less than 0.6° upon insertion of five water

molecules into the first hydration sphere. The symmetric stretching vibrational frequency

calculated with the PBE functional is lower in [PuO2(H2O)5]2+

but is still within 5.3 cm-1

of the

experimental value while the asymmetric counterpart is now about 24.5 cm-1

away from the

experimental value and the value obtained for PuO22+

. The structural features and vibrational

frequencies computed with the RECPs are in agreement44

with those obtained with the all-

electron basis set with the ZORA45-46

approach. In addition, the discrepancies between the

geometrical parameters obtained in the scalar-relativistic and spin-orbit corrected calculations are

small.44

This is in agreement with previous results and also indicates that using the small-core

pseudopotential with GGA functionals results in geometrical parameters of sufficient accuracy.31,

47 The use of the B3LYP functional with RECPs in optimizing the geometry of [PuO2(H2O)5]

2+

results in Pu=Oyl bond lengths about 0.03 Å smaller than the experimental value.35

The average

Page 95: Relativistic Quantum Chemistry Applied to Actinides

73

of the Pu-OH2 bond lengths is only slightly smaller than that obtained with the PBE functional.

The error in the calculated Pu-OH2 bond lengths is smaller than that obtained for the U-OH2

bonds in the counterpart uranyl complex, Figure 3.1. The vibrational frequency of the symmetric

stretching of the plutonyl entity in [PuO2(H2O)5]2+

obtained with the B3LYP functional is still

much larger than the experimental values. This agrees with previous experience that showed

hybrid functionals give vibrational frequencies that are generally higher than the experimental

value.10, 13, 19-20, 22

A scaling factor of 0.92-0.98 is often required to bring sufficient accordance

with the experimental data. The asymmetric O=Pu=O vibration obtained with the B3LYP

functional in the aqueous phase is not much better than that obtained with the PBE functional. A

scaling factor of about 0.93 appears sufficient in bringing a match between the experimental and

calculated vibrational frequencies obtained with the B3LYP functional. Overall however,

introduction of the five aquo ligands in the first solvation sphere only results in about 0.01-0.015

Å elongation of the Pu=Oyl bonds and a decrease in the stretching vibrational frequencies by

between 12 and 33 cm-1

.

In general, examination of the bond lengths and O=An=O stretching vibrational

frequencies calculated for the bare, AnO22+

and [AnO2(H2O)5]2+

complexes of uranium,

neptunium and plutonium indicates that the B3LYP functional generally results in slightly

shorter bonds while the PBE functional results in better O=An=O stretching vibrational

frequencies, Figure 3.1. The elongation of the An=Oyl and An-OH2 bonds upon reduction to the

+5 oxidation state is well known, Figure 3.2. The experimental values of the asymmetric

O=An=O vibration frequencies are accurately predicted by aqueous phase calculations on the

bare AnO2n+

cations with the PBE functional even though inclusion of the first hydration sphere

is necessary in reducing the differences between the experimental and calculated values of the

Page 96: Relativistic Quantum Chemistry Applied to Actinides

74

symmetric O=An=O vibrations. These observations have been previously reported.31

Overall, it

appears that the performance of DFT is in no way significantly worse for the PuO2n+

systems

than for the UO2n+

systems, Figures 3.1 and 3.2. Regarding the fortuity of the agreement between

the calculated and experimental vibrational frequencies, there is the age-old case of "getting

good results by error cancellations". At the current level of theory, the results (both the

calculated values and the trends between the calculated and experimental values) suggest that

these types of calculations can be used in future combined theoretical and experimental works.

The applicability to uranium and neptunium complexes is also advantageous.

Figure 3.1: (Left) The An=Oyl bond lengths (Å) in the bare actinyl moieties, AnO22+

as well as

the An=Oyl and average An-OH2 bond lengths (Å) in the pentaaquo complexes, [AnO2(H2O)5]2+

.

(Right) The actinyl symmetric and asymmetric stretching vibrational frequencies (cm-1

) in the

bare actinyl moieties, AnO22+

and the pentaaquo complexes, [AnO2(H2O)5]2+

. These values were

obtained with the obtained using the PBE functional and PCM solvation model.

Page 97: Relativistic Quantum Chemistry Applied to Actinides

75

Figure 3.2: (Left) The An=Oyl bond lengths (Å) in the bare actinyl moieties, AnO21+

as well as

the An=Oyl bond lengths and average An-OH2 bond lengths (Å) in the pentaaquo complexes,

[AnO2(H2O)5]1+

. (Right) The actinyl symmetric and asymmetric stretching vibrational

frequencies (cm-1

) in the bare actinyl moieties, AnO21+

and the pentaaquo complexes,

[AnO2(H2O)5]1+

. These values were obtained with the PBE functional and PCM solvation model.

Plutonyl (VI) chloro complexes. Runde et al.48

identified the mono- and bis-chloro (and a

possible tri-chloro) complexes formed with the PuO22+

moiety in sodium chloride solutions.

They found the chloro-complexes to occur at greater concentrations in such solutions than the

pentaaquo ion with the mono-chloride complex being most stable. From the X-ray absorption

spectra in solution, they and Conradson et al.35

provided the geometrical parameters of plutonyl

complexes formed at various chloride concentrations. Regarding the bis-chloride complex,

Page 98: Relativistic Quantum Chemistry Applied to Actinides

76

[PuO2Cl2(H2O)2], Berthon et al.49

recently used DFT analysis to confirm the greater stability of

the trans isomer of [PuO2Cl2(TPPO)2], TPPO=triphenylphosphine oxide. This structure is

analogous to trans-[PuO2Cl2(H2O)2] which our solution phase calculations with the B3LYP

functional shows to be essentially iso-energetic with its cis-isomer. The energy difference,

ΔG(cis-trans), between the geometrical isomers of [PuO2Cl2(H2O)2] was calculated as 3.5 and

0.0 kcal/mol when the B3LYP functional was employed in the gaseous and aqueous phases,

respectively. Single-point calculation with the aug-cc-pVTZ basis (for O, Cl and H atoms) on the

geometries optimized at the double-ζ level reduces the cis-trans energy difference to about 0.0

kcal/mol in the aqueous phase. The magnitude of the energy differences calculated suggests that

both structures are iso-energetic with ready inter-conversion in aqueous solutions. The greater

stability of the trans- structure of [PuO2Cl2(TPPO)2] is most likely due to steric repulsion.

Structurally, The Pu=Oyl, Pu-OH2 and Pu-Cl bond lengths in the trans-isomer of

[PuO2Cl2(H2O)2] were calculated as 1.736, 2.417 and 2.643 Å, respectively, in the aqueous phase

with the B3LYP functional, Table 3.2. The Pu-OH2 bonds are slightly longer (2.444 Å) in the

cis-isomer. The calculated Pu=Oyl bond lengths are in agreement with the experimental value of

1.75 Å which is relatively unchanged from those in the pentaaquo complex. The calculated Pu-

OH2 and Pu-Cl bonds are however significantly shorter than the values of 2.49 and 2.70 Å

respectively reported by Conradson et al.35

and Runde et al.48

It should however be noted that the

calculated Pu-Cl bonds obtained for the [PuO2Cl2(H2O)2] complexes in this work are in good

agreement with those obtained for the PuO2Cl2L2 complexes from DFT calculations (2.634-

Page 99: Relativistic Quantum Chemistry Applied to Actinides

77

Table 3.2: The calculated properties (bond lengths in Å and frequencies in cm-1

) of plutonyl

chloro-aquo complexes obtained using the B3LYP (and PBE) functional.a W is used to represent

the aquo ligand (H2O).

Complexb Gas Phase Aqueous Solution

Oyl Aquo Chloro υsymm/asymm Oyl Aquo Chloro υsymm/asymm

PuO2Cl42-

1.74

(1.77)

2.73

(2.71)

824/941

(763/878)

1.75

(1.78)

2.70

(2.67)

816/917

(756/858)

PuO2Cl3W1 1.74

(1.77)

2.59

(2.62)

2.67

845/961

(776/898)

1.74

(1.77)

2.52

(2.53)

2.68

(2.67)

833/934

(769/877)

PuO2Cl2W2

Cis- 1.74

(1.77)

2.51

(2.52)

2.59

(2.58)

852/967

(800/914)

1.74 2.44 2.64 848/946

Trans- 1.74

(1.76)

2.48

(2.51)

2.60

(2.60)

850/967

(795/915)

1.74

(1.77)

2.42

(2.42)

2.64

(2.62)

845/945

(786/891)

PuO2Cl2W3

Cis- 1.73

(1.77)

2.62

(2.62)

2.63

(2.63)

859/973

(785/909)

1.73

(1.77)

2.53

(2.55)

2.67

(2.66)

847/947

(780/890)

Page 100: Relativistic Quantum Chemistry Applied to Actinides

78

Trans- 1.73

(1.76)

2.55

(2.55)

2.67

(2.67)

864/980

(798/919)

1.73

(1.76)

2.53

(2.53)

2.69

(2.68)

848/948

(784/892)

Expt.c 1.75 2.49 2.70

PuO2Cl1W4 1.73

(1.76)

2.53

(2.54)

2.58

(2.57)

879/995

(813/934)

1.73

(1.76)

2.48

(2.50)

2.64

(2.62)

867/965

(802/910)

Expt.c 1.75 2.43 2.75

a The calculated Pu-Oyl, Pu-OH2 and Pu-Cl bond lengths are given.

b W is used to represent the

aquo ligand (H2O). c Reference

48

2.637 Å)49

and experimental measurements (2.630 Å).48

Interestingly, addition of another aquo

ligand to the equatorial region results in elongation of the Pu-Cl bonds to 2.679 and 2.689 Å in

the cis- and trans- isomers of [PuO2Cl2(H2O)3] respectively, Table 3.2 and Figure 3.3. This is in

much better agreement with the measurements of Runde et al.48

The Pu-OH2 bonds also increase

to 2.534 and 2.522 Å in the cis- and trans- isomers respectively in much better agreement with

the EXAFS value of 2.49 Å.48

The calculated structural features thus allow us to identify the

previously characterized aqueous phase complexes characterized as [PuO2Cl2(H2O)3] rather than

[PuO2Cl2(H2O)2] for which there are many analogous PuO2Cl2L2 solid-state species. Such

structural differences between the di-aquo and tri-aquo species can also be observed in the bis-

chloro complexes formed with the uranyl group.2 For [PuO2Cl2(H2O)3], the trans- structure is

about 5.9 and 1.6 kcal/mol more stable than the cis- structure in the gas and aqueous phase

calculations with the B3LYP functional respectively. This larger energy difference implies the

Page 101: Relativistic Quantum Chemistry Applied to Actinides

79

trans- isomer should be significantly predominant in aqueous solutions. It can be directly related

to the increased steric crowding compared to [PuO2Cl2(H2O)2]. Also, the structural features of

the trans- complex are in slightly better agreement with the experimental reports, Table 3.2.

The O=Pu=O asymmetric stretching vibration has been observed between 908 to 918 cm-

1 in previous solid-state measurements of the IR spectra of various PuO2Cl2L2-type complexes.

50

Calculations with the PBE functional on [PuO2Cl2(H2O)2] in the gas-phase predict the

asymmetric stretching vibration frequency of O=Pu=O to be 914.9 and 914.4 cm-1

for the trans-

and cis- structures respectively, Table 3.2. It should be noted that the gas-phase calculations also

predict this vibrational frequency mode to be at 971.0 cm-1

in [PuO2(H2O)5]2+

. The agreement

between the values calculated for the PuO2Cl2L2 and [PuO2Cl2(H2O)2] complexes indicates

considerable decrease in the plutonyl vibrational frequencies upon introduction of the chloride

ligands into the equatorial region. It appears that the role of the organic ligand is minimal in this

effect. Indeed the asymmetric vibrational mode of the plutonyl group in trans-[PuO2Cl2(TPPO)2]

was calculated at 916.7 cm-1

in the gas-phase using the PBE functional, adequately agreeing with

the peak at 920 cm-1

observed in the solid state IR spectrum.

In the [PuO2Cl2(H2O)3] complex, the asymmetric O=Pu=O stretching vibration is

between 908.7 and 919.3 cm-1

in the gas phase and between 889.7 and 892.4 cm-1

in the aqueous

phase, Table 3.2. It appears that the magnitude of the O=Pu=O asymmetric stretching vibration

frequency is much more dependent on the introduction of two chloride ligands into the equatorial

region and less so on the number of aquo ligands in the complex. The symmetric vibrational

frequencies of the plutonyl moiety in [PuO2Cl2(H2O)3] were calculated to be between 780.3-

784.4 cm-1

in the aqueous phase, a drastic reduction from the 829.7 cm-1

(and 835 cm-1

)

calculated (and experimentally measured) for the pentaaquo complex. The reduction of both the

Page 102: Relativistic Quantum Chemistry Applied to Actinides

80

symmetric and asymmetric stretching vibrations of the O=Pu=O group by 45 to 60 cm-1

upon

introduction of two chloride ligands into the equatorial region is supported by the theoretical

calculations in this work and previously reported experimental results.50

Figure 3.3: A) The cis (top) and trans (bottom) isomers of the PuO2Cl2(H2O)2 complex. B) The

cis (top) and trans (bottom) isomers of the PuO2Cl2(H2O)3 complex.

The mono-chloro complexes formed at lower concentrations were calculated as having

shorter Pu-OH2 and Pu=Oyl bonds than the bis-chloro complexes, Table 3.2. This is in agreement

with the experimental reports. The Pu-Cl bond length in [PuO2Cl(H2O)4]+ was calculated as

2.642 Å in the aqueous solution significantly underestimating the experimental value of 2.75 Å

for the Pu-Cl bonds in the complex formed by PuO22+

at lower chloride concentrations in acidic

aqueous media. The experimental report of much longer (by about 0.05 Å) Pu-Cl bonds in the

Page 103: Relativistic Quantum Chemistry Applied to Actinides

81

complex formed at lower chloride concentrations is surprising as similar effects are absent in the

uranyl counterparts.51-52

Indeed after replacement of all the aquo ligands by chloride anions, the

Pu-Cl bonds in [PuO2Cl4]2-

were calculated to be only about 2.697 Å in length in aqueous

solution. The calculated Pu-Cl bond lengths of [PuO2Cl4]2-

in the gas and aqueous phases are in

agreement with the theoretical results of Austin et al.31

They are also near the range of 2.653-

2.671 Å previously reported in experimental works on crystalline Cs2[AnO2Cl4] complexes.53-55

Figure 3.4: A) The decrease in the An=Oyl and An-Cl bond lengths (Å) on addition of 2Cs+ to

the [AnO2Cl4]2-

complexes. B) The increase in the actinyl stretching vibrational frequencies (cm-

1) on addition of 2Cs

+ to the [AnO2Cl4]

2- complexes. The structural features of the anionic and

neutral complexes were calculated with the PBE functional and are depicted with black and red

squares respectively. The experimental values56, 59

of the symmetric vibrational frequencies in

solid Cs2[UO2Cl4] and Cs2[NpO2Cl4] are 832 and 802 cm-1

respectively. The calculated values

for these complexes are within 17 cm-1

and 4 cm-1

of the experimental reports respectively.

Page 104: Relativistic Quantum Chemistry Applied to Actinides

82

For the [AnO2Cl4]2-

complexes in general, Figure 3.4, the trends in the calculated bond lengths

and vibrational frequencies in the gas phase are similar to those observed in the pentaaquo

uranyl, neptunyl and plutonyl complexes, Figure 3.1. The attachment of cesium atoms to two

chloride ligands in a trans- arrangement allows us to simulate the single-crystal units of the

Cs2[AnO2Cl4] complexes. The calculations show the cesium salts should have shorter An-Cl

bonds (by about 0.09 Å for the non-cesiated chloride ligands although the ones with Cs salts

maintain the same bond lengths found in the [AnO2Cl4]2-

species) than the counterpart

[AnO2Cl4]2-

complexes, Figure 3.4. The difference between the An=Oyl bonds in the gas phase

anionic and neutral salts are more modest (less than 0.01 Å). Overall these changes upon the

addition of two cesium ions bring the An-Cl and An=Oyl bonds in agreement with available solid

state experimental53-55

reports on the actinyl complexes. The symmetric and asymmetric

stretching vibrational frequencies of the actinyl entities are larger in the cesium complexes,

Figure 3.4. The calculated values of the symmetric and asymmetric vibrations of actinyl entity in

the cesiated neptunyl and uranyl complexes are in good agreement with the values reported by

Denning.56-59

Plutonyl (VI) nitrate complexes. Gaunt et al.32

recently crystallized plutonyl (VI) diaquo-

dinitrate and characterized it using X-ray diffraction techniques. The crystalline dinitrate they

obtained had two water molecules in a trans- arrangement in the equatorial region. They however

observed only negligible dinitrate formation in aqueous solutions even at very low pH values.

The aqueous solution of PuO22+

in acidic nitrate solutions shows very weak coordination of the

nitrate ligand with the mononitrate complex being the dominant species. The calculated

structural parameters of a few structures of the mononitrate, dinitrate and trinitrate complexes are

given in Table 3.3. For the diaquo-dinitrate characterized by Gaunt et al.32

, Figure 3.5, the

Page 105: Relativistic Quantum Chemistry Applied to Actinides

83

experimental Pu=Oyl bond length of 1.727 Å is well replicated by the B3LYP/RECPs

calculations as 1.728 Å. The calculated Pu-OH2 bond lengths at 2.534 Å are however

significantly larger than the experimental values, 2.450 Å. Such a large discrepancy was also

seen between the calculated and experimental U-OH2 bonds in the work on the analogous U(VI)

complex by Gutowski et al.6 and Prestianni et al.

16 The Hartree-Fock calculations on the uranyl

nitrate and sulfate complexes by Craw et al.60

also yielded this over-estimation of the U-OH2

bond lengths. The bonds between the actinide atom and the coordinating oxygen atoms of the

two nitrate groups in [PuO2(NO3)2(H2O)2] were calculated to be about 2.469 Å in length, a value

near the average value of 2.497 Å obtained in the X-ray diffraction study. It should be noted that

the unit cell of the dinitrate in the experimental study contains two independent molecules. There

appears to be a mismatch between the calculated and experimental relative magnitudes of the Pu-

Onitrate and Pu-OH2 bond lengths due to the significant overestimation of the latter.

The symmetric O=Pu=O and coordinated nitrate stretching vibrational frequencies of the

dinitrate are Raman active and were measured at 844 and 756 cm-1

, respectively in the solid

phase.32

The calculated infrared values for these frequencies are 866.9 and 741.9 cm-1

,

respectively at the B3LYP/RECP level in the gas phase, Table 3.3. The experimental value of the

plutonyl symmetric stretching vibrational frequency is over-estimated by approximately 23 cm-1

at this level. The calculated values for these vibrations in the aqueous phase are 847.7 and 748.4

cm-1

respectively. This suggests that, if the diaquo-dinitrate complex is formed in solution, the

symmetric stretching frequency of the plutonyl entity should be lower than that obtained in the

gas phase and also lower than the experimental value of this vibrational mode in the [PuO2

(H2O)5]2+

complex.61

The absence of any peaks around 820-830 cm-1

in the Raman spectrum of

Page 106: Relativistic Quantum Chemistry Applied to Actinides

84

Table 3.3: The calculated structural properties (bond lengths in Å and vibrational frequencies in cm-1

) of the plutonyl nitrate

complexes obtained using the B3LYP functional in the gaseous and aqueous phases.

Gas Phase Aqueous Solution

Oyl Aquo Nitrate υsymm/asymm υnitrate Oyl Aquo Nitrate υsymm/asymm υnitrate

[PuO2(NO3)]1-

1.708 2.364 911.4/1001.0 669.3 1.719 2.437 871.4/980.1 698.8

[PuO2(NO3)W4]1+

1.733 2.592 2.420 867.2/993.1 746.6 1.733 2.552 2.486 863.0/970.1 748.6

[PuO2(NO3)2W2] 1.740 2.531 2.462 850.2/976.1 746.5 1.743 2.512 2.477 847.7/948.6 748.4

Expt.a 1.727 2.432 2.497 844 756

[PuO2(NO3)3]1-

1.738 2.487 850.1/965.3 747.7 1.737 2.483 845.9/943.1 745.8

a Reference

32

Page 107: Relativistic Quantum Chemistry Applied to Actinides

85

Figure 3.5: The optimized structure of plutonyl diaquo-dinitrate, PuO2(NO3)2(H2O)2 obtained

using the B3LYP functional.

the plutonyl moiety in acidic nitrate solutions most likely suggests negligible formation of the

dinitrate in agreement with the conclusion of Gaunt et al.32

The peak observed at 836 cm-1

in the

aqueous solution of PuO22+

in concentrated nitric acid corresponds to the symmetric stretching

frequency in the pentaaquo complex.

The nitrate counterparts of the [PuO2Cl2L2] complexes, [PuO2(NO3)2L2], are of great

importance in the nuclear processing industry. In summary organic ligands are used in liquid-

liquid extraction of actinide compounds from complex acidic solutions. Plutonyl nitrate is

extracted with tributylphosphate, TBP, in the PUREX extraction process.62

The structural

parameters of the [AnO2(NO3)2(H2O)2] and [AnO2(NO3)2(TBP)2] complexes (An = U, Np and

Pu), optimized in the gas phase are presented in Table 3.4. Overall, there appears to be quite little

change in the An=Oyl bond lengths between the corresponding [AnO2(NO3)2(H2O)2] and

Page 108: Relativistic Quantum Chemistry Applied to Actinides

86

Table 3.4: The calculated bond lengths (Å) of the AnO2(NO3)2(H2O)2, AnO2(NO3)2(TBP)2a,

AnO2Cl2(H2O)2 and AnO2Cl2(TPPO)2a complexes obtained using the B3LYP functional with

RECPs in the gaseous phase

AnO2(NO3)2(H2O)2 AnO2(NO3)2(TBP)2a,b

U Np Pu U Np Pu

An=Oyl Calc. 1.769 1.748 1.740 1.768 1.749 1.743

Expt. 1.754-1.763 1.727 1.77 1.75 1.75

An-Onitrate Calc. 2.483 2.469 2.462 2.516 2.496 2.489

Expt. 2.477-2.513 2.497 2.54 2.51 2.50

An-OH2/OP Calc. 2.546 2.546 2.531 2.428 2.385 2.443

Expt. 2.446-2.457 2.432 2.41 2.38 2.40

AnO2Cl2(H2O)2 AnO2Cl2(TPPO)2a,c

An=Oyl Calc. 1.773 1.754 1.737 1.777 1.763 1.751

Expt. 1.753-1.767 1.751 1.747

An-Cl Calc. 2.622 2.604 2.597 2.650 2.627 2.629

Expt. 2.645-2.673 2.622 2.630

An-OH2/OP Calc. 2.509 2.494 2.478 2.352 2.330 2.326

Page 109: Relativistic Quantum Chemistry Applied to Actinides

87

Expt. 2.300-2.339 2.288 2.302

a TPPO and TBP are abbreviations for triphenylphosphine oxide and tributylphosphate

respectively. b Reference

63.

c Reference

49, 65-67

[AnO2(NO3)2(TBP)2] complexes. The structural features of the uranyl, neptunyl and plutonyl

complexes are all rather similar with minimal change across the series, Table 3.4. This as well as

the calculated bond distances are in good agreement with the experimental reports of den Auwer

et al.63

This is also the case in the [AnO2Cl2(TPPO)2] complexes, Table 3.4. Due to

electronegativity differences, the An-Ophosphine bonds in the [AnO2(NO3)2(TBP)2] complexes are

shorter than the An-OH2 bonds in their [AnO2(NO3)2(H2O)2] counterparts, Table 3.4. This is also

evident in the dichloro- complexes. This goes back to the ionic nature of the equatorial bonds in

actinyl complexes.

Plutonyl (IV) aquo and nitrate complexes. The ability of DFT to successfully predict the

structural features of Pu(IV) complexes has been demonstrated recently by Horowitz et al.29

The

Pu-OH2 bond length in the [Pu(H2O)8]4+

complex was calculated as 2.39 Å in good agreement

with the experimental value of 2.38 Å. Our calculations in the aqueous phase yield Pu-OH2 bond

lengths of 2.392 and 2.393 Å with the B3LYP and PBE functionals respectively, Table 3.5. The

dinitrate, tetranitrate and hexanitrate complexes of Pu4+

are known to be the dominant species in

acidic aqueous nitrate solutions. The pentanitrate and hexanitrate are thought to be present at

high nitrate concentrations in acidic aqueous solutions. The geometries of the [Pu(NO3)n]4-n

(n=2-6) complexes were optimized using both density functionals employed in this work. The

Page 110: Relativistic Quantum Chemistry Applied to Actinides

88

Table 3.5: The calculated Pu-OH2 and Pu-Owater/nitrate bond lengths of the aquo, nitrate and aquo-

nitrate complexes of the plutonium (IV) cation obtained using DFT in the gaseous and aqueous

phases

PBE B3LYP Expt.a

Gas Solution Gas Solution

[Pu(H2O)8]4+

2.428 2.393 2.416 2.393 2.39

[Pu(NO3)2]2+

2.231 2.419 2.214 2.390 „2.42‟

[Pu(NO3)3]1+

2.304 2.450 2.298 2.411

[Pu(NO3)4]0 2.385 2.438 2.376 2.434

[Pu(NO3)4(H2O)3]0 2.504 2.511 2.500 2.46

[Pu(NO3)5]1-

2.456 2.474 2.449 2.469

[Pu(NO3)5(H2O)]1-

2.509 2.507 2.502

[Pu(NO3)6]2-

2.534 2.524 2.529 2.520 2.48

a References

34-35

nitrate ligands were coordinated to the Pu4+

cation in a bidentate fashion, Figure 3.6. It should be

noted that the EXAFS experiments of Allen et al.34

found average Pu-Owater/nitrate bond lengths of

2.42, 2.46 and 2.49 Å respectively for the heptaaquo-dinitrate, triaquo-tetranitrate and

hexanitrate species. The average of the calculated Pu-Onitrate bond lengths in the dinitrate,

Page 111: Relativistic Quantum Chemistry Applied to Actinides

89

tetranitrate and hexanitrate species obtained with the B3LYP functional in aqueous solution are

2.399, 2.438 and 2.520 Å respectively, Table 3.5. This indicates that in the absence of aquo

ligands, the Pu-Onitrate bond lengths can vary between 2.399 and 2.520 Å. The Pu-Owater bonds

would then be expected to slightly increase the average Pu-Owater/nitrate bond lengths in the

[Pu(NO3)n(H2O)11-2n]4-n

complexes. Indeed optimization of the [Pu(NO3)4(H2O)3] and

[Pu(NO3)5(H2O)1]1-

aquo-nitrate complexes results in longer (0.03-0.07 Å) average Pu-

Owater/nitrate bonds.

Figure 3.6: The optimized structures of several plutonium (IV) complexes: A) Pu(H2O)84+

, B)

Pu(NO3)22+

, C) Pu(NO3)4(H2O)3 and D) Pu(NO3)62-

calculated with the B3LYP functional in the

aqueous phase.

Page 112: Relativistic Quantum Chemistry Applied to Actinides

90

Conclusions

We have performed a comprehensive characterization of the structural properties of

several plutonyl (VI) and plutonium (IV) complexes using scalar relativistic Kohn-Sham DFT

calculations with small-core effective core pseudopotentials. The aquo, chloro and nitrate

complexes were examined in the gas and aqueous phases in addition to the complexes formed

with a few organic ligands. The calculations in aqueous solvents were carried out using the PCM

solvation model.

Overall, good performance can be obtained by using the hybrid B3LYP functional in

calculating the structural parameters of the plutonyl (VI) and plutonium (IV) complexes or by

using the PBE functional in predicting the signature infrared and Raman active vibrational

frequencies of these complexes either in the gas-phase (solid-state single crystals) or aqueous

phases. The use of the vibrational frequencies predicted by these functionals in addition with

experimental infrared or Raman spectroscopic data promises to be a strong tool in confirming the

identity of plutonium complexes that might be isolated or prepared in the future. It should be

noted that the calculated vibrational frequencies rely on the harmonic approximation. Deviations

between the experimental and calculated vibrational frequencies can therefore be expected. The

agreement between the two is however sufficient for future combined theoretical and

experimental work on actinide complexes. In terms of geometrical parameters, the plutonyl

complexes are rather very reminiscent of their counterpart uranyl and neptunyl counterparts with

usually minimal changes in the associated bond lengths (0.01-0.03 Å) and bond angles. The only

problematic point seems to be the calculation of the An-OH2 bond lengths in the diaquo-dinitrate

complexes. It might be necessary to explicitly include water molecules in the second

coordination sphere around the actinyl entity for good agreement between the experimental and

Page 113: Relativistic Quantum Chemistry Applied to Actinides

91

calculated An-OH2 bond lengths. We have extended the use of DFT in predicting the structure of

plutonium (IV) complexes beyond the octaaquo complexes. The calculated and experimental

geometrical parameters in the nitrate and aquo-nitrate complexes agree to within 0.02-0.04 Å.

It would appear that DFT calculations can be used in predicting sufficiently accurate

structural parameters for such complexes, the lack of rigor in not fully accounting for the various

nearly-degenerate electronic states (correct ground state?) and their multi-reference characters

notwithstanding. The scope of the current work is limited to examining the structural features of

the selected molecules with limited examination of the ability of DFT to correctly predict the

speciation, coordination numbers and hydration energies of these plutonium compounds. Going

forward, the use of spin-orbit coupled wave function approaches in predicting the electronic

structure and spectroscopic properties of small to medium-sized actinide species holds some

appeal for us.

References

1. Batista, E. R.; Martin, R. L.; Hay, P. J., J. Chem. Phys. 2004, 121, 11104.

2. Bühl, M.; Sieffert, N.; Golubnychiy, V.; Wipff, G., J. Phys. Chem. A. 2008, 112, 2428.

3. Clavaguera-Sarrio, C.; Ismail, N.; Marsden, C. J.; Begue, D.; Pouchan, C., Chem. Phys.

2004, 302, 1.

4. Gagliardi, L.; Willetts, A.; Skylaris, C. K.; Handy, N. C.; Spencer, S.; Ioannou, A. G.;

Simper, A. M., J. Am. Chem. Soc. 1998, 120, 11727.

5. Garcia-Hernandez, M.; Lauterbach, C.; Krüger, S.; Matveev, A.; Rösch, N., J. Comput.

Chem. 2002, 23, 834.

Page 114: Relativistic Quantum Chemistry Applied to Actinides

92

6. Gutowski, K. E.; Cocalia, V. A.; Griffin, S. T.; Bridges, N. J.; Dixon, D. A.; Rogers, R.

D., J. Am. Chem. Soc. 2007, 129, 526.

7. Han, Y. K., J. Comput. Chem. 2001, 22, 2010.

8. Han, Y. K.; Hirao, K., J. Chem. Phys. 2000, 113, 7345.

9. Hay, P. J.; Martin, R. L., J. Chem. Phys. 1998, 109, 3875.

10. Iche-Tarrat, N.; Marsden, C. J., J. Phys. Chem. A. 2008, 112, 7632.

11. Kaltsoyannis, N.; Hay, P. J.; Li, J.; Blaudeau, J. P.; Bursten, B. E., Theoretical Studies of

the Electronic Structure of Compounds of the Actinides. In The Chemistry of the Actinide and

Transactinide Elements, 3rd ed, 3 ed.; Morss, L. R.; Edelstein, N. M.; Fuger, J.; Katz, J. J., Eds.

Springer: Dordrecht, The Netherlands, 2006; Vol. 3, pp 1893.

12. Kovács, A.; Konings, R. J. M., J. Mol. Struct-Theochem 2004, 684, 35.

13. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2010, 114, 1957.

14. Peralta, J. E.; Batista, E. R.; Scuseria, G. E.; Martin, R. L., J. Chem. Theory Comput.

2005, 1, 612.

15. Perron, H.; Roques, J.; Domain, C.; Drot, R.; Simoni, E.; Catalette, H., Inorg. Chem.

2008, 47, 10991.

16. Prestianni, A.; Joubert, L.; Chagnes, A.; Cote, G.; Ohnet, M. N.; Rabbe, C.; Charbonnel,

M. C.; Adamo, C., J. Phys. Chem. A. 2010, 114, 10878.

17. Schimmelpfennig, B.; Privalov, T.; Wahlgren, U.; Grenthe, I., J. Phys. Chem. A. 2003,

107, 9705.

18. Schreckenbach, G.; Hay, P. J.; Martin, R. L., J. Comput. Chem. 1999, 20, 70.

19. Schreckenbach, G.; Shamov, G. A., Acc. Chem. Res. 2010, 43, 19.

20. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109, 10961.

Page 115: Relativistic Quantum Chemistry Applied to Actinides

93

21. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A. 2006, 110, 9486.

22. Shamov, G. A.; Schreckenbach, G.; Vo, T. N., Chem-Eur. J. 2007, 13, 4932.

23. Vallet, V.; Macak, P.; Wahlgren, U.; Grenthe, I., Theor. Chem. Acc. 2006, 115, 145.

24. Vallet, V.; Wahlgren, U.; Schimmelpfennig, B.; Moll, H.; Szabo, Z.; Grenthe, I., Inorg.

Chem. 2001, 40, 3516.

25. Cao, Z. J.; Balasubramanian, K., J. Chem. Phys. 2005, 123.

26. Chaudhuri, D.; Balasubramanian, K., Chem. Phys. Lett. 2004, 399, 67.

27. Clavaguera-Sarrio, C.; Vallet, V.; Maynau, D.; Marsden, C. J., J. Chem. Phys. 2004, 121,

5312.

28. Ismail, N.; Heully, J. L.; Saue, T.; Daudey, J. P.; Marsden, C. J., Chem. Phys. Lett. 1999,

300, 296.

29. Horowitz, S. E.; Marston, J. B., J. Chem. Phys. 2011, 134.

30. La Macchia, G.; Infante, I.; Raab, J.; Gibson, J. K.; Gagliardi, L., Phys. Chem. Chem.

Phys. 2008, 10, 7278.

31. Austin, J. P.; Sundararajan, M.; Vincent, M. A.; Hillier, I. H., Dalton T. 2009, 5902.

32. Gaunt, A. J.; May, I.; Neu, M. P.; Reilly, S. D.; Scott, B. L., Inorg. Chem. 2011, 50,

4244.

33. Chaudhuri, D.; Balasubramanian, K., Chem. Phys. Lett. 2004, 399, 67.

34. Allen, P. G.; Veirs, D. K.; Conradson, S. D.; Smith, C. A.; Marsh, S. F., Inorg. Chem.

1996, 35, 2841.

35. Conradson, S. D.; Abney, K. D.; Begg, B. D.; Brady, E. D.; Clark, D. L.; den Auwer, C.;

Ding, M.; Dorhout, P. K.; Espinosa-Faller, F. J.; Gordon, P. L.; Haire, R. G.; Hess, N. J.; Hess,

R. F.; Keogh, D. W.; Lander, G. H.; Lupinetti, A. J.; Morales, L. A.; Neu, M. P.; Palmer, P. D.;

Page 116: Relativistic Quantum Chemistry Applied to Actinides

94

Paviet-Hartmann, P.; Reilly, S. D.; Runde, W. H.; Tait, C. D.; Veirs, D. K.; Wastin, F., Inorg.

Chem. 2004, 43, 116.

36. Frisch, M. J. T., G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.;

Montgomery, Jr., J. A.; Vreven, T.; Kudin, K. N.; Burant, J. C.; Millam, J. M.; Iyengar, S. S.;

Tomasi, J.; Barone, V.; Mennucci, B.; Cossi, M.; Scalmani, G.; Rega, N.; Petersson, G. A.;

Nakatsuji, H.; Hada, M.; Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Klene, M.; Li, X.; Knox, J. E.; Hratchian, H. P.; Cross, J. B.;

Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;

Cammi, R.; Pomelli, C.; Ochterski, J. W.; Ayala, P. Y.; Morokuma, K.; Voth, G. A.; Salvador,

P.; Dannenberg, J. J.; Zakrzewski, V. G.; Dapprich, S.; Daniels, A. D.; Strain, M. C.; Farkas, O.;

Malick, D. K.; Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Ortiz, J. V.; Cui, Q.; Baboul,

A. G.; Clifford, S.; Cioslowski, J.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.;

Komaromi, I.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,

A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen, W.; Wong, M. W.; Gonzalez, C.; and

Pople, J. A. Gaussian 03, Revision C.02. 2004.

37. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74, 1245.

38. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100, 7535.

39. Becke, A. D., J. Chem. Phys. 1993, 98, 5648.

40. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J., J. Phys. Chem. 1994, 98,

11623.

41. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1996, 77, 3865.

42. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1997, 78, 1396.

43. Miertus, S.; Scrocco, E.; Tomasi, J., Chem. Phys. 1981, 55, 117.

Page 117: Relativistic Quantum Chemistry Applied to Actinides

95

44. Odoh, S. O., Unpublished results.

45. van Lenthe, E., J. Comput. Chem. 1999, 20, 51.

46. van Lenthe, E.; Baerends, E. J.; Snijders, J. G., J. Chem. Phys. 1993, 99, 4597.

47. Odoh, S. O.; Walker, S. M.; Meier, M.; Stetefeld, J.; Schreckenbach, G., Inorg. Chem.

2011, 50, 3141.

48. Runde, W.; Reilly, S. D.; Neu, M. P., Geochim. Cosmochim. Ac. 1999, 63, 3443.

49. Berthon, C.; Boubals, N.; Charushnikova, I. A.; Collison, D.; Cornet, S. M.; Auwer, C.;

Gaunt, A. J.; Kaltsoyannis, N.; May, I.; Petit, S.; Redmond, M. P.; Reilly, S. D.; Scott, B. L.,

Inorg. Chem. 2010, 49, 9554.

50. Balakris.P. V; Patil, S. K.; Sharma, H. D.; Venkatas.H. V, Can. J. Chemistry 1965, 43,

2052.

51. Allen, P. G.; Bucher, J. J.; Shuh, D. K.; Edelstein, N. M.; Reich, T., Inorg. Chem. 1997,

36, 4676.

52. Hennig, C.; Tutschku, J.; Rossberg, A.; Bernhard, G.; Scheinost, A. C., Inorg. Chem.

2005, 44, 6655.

53. Watkin, D. J.; Denning, R. G.; Prout, K., Acta Crystallogr. C 1991, 47, 2517.

54. Wilkerson, M. P.; Arrington, C. A.; Berg, J. M.; Scott, B. L., J. Alloy Compd. 2007, 444,

634.

55. Wilkerson, M. P.; Scott, B. L., Acta Crystallogr. E 2008, 64, I5.

56. Denning, R. G., Struct. Bond. 1992, 79, 215.

57. Denning, R. G., J. Phys. Chem. A. 2007, 111, 4125.

58. Denning, R. G.; Morrison, I. D., Chem. Phys. Lett. 1991, 180, 101.

59. Denning, R. G.; Norris, J. O. W.; Brown, D., Mol. Phys. 1982, 46, 325.

Page 118: Relativistic Quantum Chemistry Applied to Actinides

96

60. Craw, J. S.; Vincent, M. A.; Hillier, I. H.; Wallwork, A. L., J. Phys. Chem. 1995, 99,

10181.

61. Madic, C.; Begun, G. M.; Hobart, D. E.; Hahn, R. L., Inorg. Chem. 1984, 23, 1914.

62. http://www.ne.doe.gov/pdfFiles/NRCseminarreprocessing_Terry_Todd.pdf.

63. Den Auwer, C.; Revel, R.; Charbonnel, M. C.; Presson, M. T.; Conradson, S. D.; Simoni,

E.; Le Du, J. F.; Madic, C., J. Synchrotron Radiat. 1999, 6, 101.

64. Basile, L. J.; Sullivan, J. C.; Ferraro, J. R.; Labonvil.P, Appl. Spectrosc. 1974, 28, 142.

65. Akona, S. B.; Fawcett, J.; Holloway, J. H.; Russell, D. R.; Leban, I., Acta Crystallogr. C

1991, 47, 45.

66. Alcock, N. W.; Roberts, M. M.; Brown, D., J. Chem. Soc.-Dalton Trans. 1982, 25.

67. Bombieri, G.; Brown, D.; Graziani, R., J. Chem. Soc.-Dalton Trans. 1975, 1873.

Page 119: Relativistic Quantum Chemistry Applied to Actinides

97

Preface to Chapter 4

This chapter is based on a manuscript published in the journal “Inorganic Chemistry”. The full

citation of the paper is as follows:

Samuel O. Odoh, Justine A. Reyes and Georg Schreckenbach, “Theoretical Study of the

Structural and Electronic Properties of Plutonyl Hydroxides”, Inorg. Chem. 2012,

submitted.

All the calculations in the main body of the published manuscript and compiled in this chapter

were carried out by Samuel O. Odoh. The manuscript was prepared together with Justine A.

Reyes and Prof. Georg Schreckenbach. Justine A. Reyes was an undergraduate research fellow in

the Department of Chemistry, University of Manitoba.

A crucial difference between the speciation of hexavalent uranyl and plutonyl complexes in

aqueous solutions at high pH has been reported. The uranyl speciation system contains a

significant amount of the trimeric (and heavier) complex at modest to high pH values. This

contrasts with the plutonyl system in which little or negligible amounts of the trimer complex

was observed. In this chapter the structure and electronic properties of plutonyl aquo-hydroxo

complexes are examined using density functional theory. The origin of the differences in the

mole-fractions of the uranyl and plutonyl trimeric complexes in the actinyl aquo-hydroxo

speciation diagrams was also determined.

Copyright permissions have been obtained from the American Chemical Society and the other

authors.

Page 120: Relativistic Quantum Chemistry Applied to Actinides

98

Chapter 4: Theoretical Study of the Structural and Electronic

Properties of Plutonyl Hydroxides

Abstract

Previous experimental reports have revealed major differences between the uranyl and plutonyl

hydrolysis systems. The uranyl hydrolysis is dominated by trimeric and polynuclear species

while the plutonyl system is dominated by a dimer complex with relatively low mole-fractions

for the trimer. To fully understand the origins of this discrepancy, the structural and electronic

properties of the monomeric, dimeric and trimeric plutonyl hydroxo complexes as well as the

hydrolysis reactions of [PuO2(H2O)4]2+

have been theoretically examined using scalar-relativistic

density functional theory. The stabilities of the plutonyl hydroxo complexes were also compared

to those of their uranyl and neptunyl analogues. The trends in the structures of the plutonyl

hydroxo complexes are generally similar to those observed in their uranyl counterparts. The

calculated plutonyl symmetric stretching vibrational frequencies were used to match previously

reported experimental Raman peaks at 833, 817, 826 and 794 cm-1

to [PuO2(H2O)5]2+

,

[PuO2(H2O)4(OH)]+, [(PuO2)2(OH)2(H2O)6]

2+ and [(PuO2)2(OH)4(H2O)4] respectively. The

assignment of the experimental peak at 805 cm-1

is equivocal as it could be assigned to either

[(PuO2)2(OH)4(H2O)4] or [(PuO2)3(H2O)6(O)(OH)3]+. The calculated reaction energies, ΔEreaction,

for the hydrolysis of [AnO2(H2O)5]2+

complexes show that the formation of the dihydroxo dimer

is significantly more exothermic than the formation of the trimer. The formation of the uranyl

trimer complex by hydrolysis of [UO2(H2O)5]2+

is however about 6.2 kcal/mol more exothermic

than the case in the plutonyl system. This calculated ΔΔEreaction agrees very well with

Page 121: Relativistic Quantum Chemistry Applied to Actinides

99

experimental estimates of 6.32 kcal/mol and could explain the dominance of the trimer in the

uranyl system in contrast to its minor role in the plutonyl hydrolysis system. The greater

oxophilicity of the uranyl group results in greater stabilization of its trimeric complex, in

comparison to that of plutonyl, by dehydration of two hydroxo groups to form a μ3-oxo group.

The presence of a bridging aquo ligand in the trimer ring also destabilizes the uranyl system to a

far greater extent than its plutonyl analog. These critical differences are manifested in the μ3-oxo

atomic 2p contributions to the actinyl-π(d)/μ3O-2p orbitals as well as the extent to which the

π(f/d)/σ(Pu-μ3O) orbitals are stabilized. These two factors decrease down the U, Np and Pu

series.

Introduction

The +6 oxidation state of plutonium is dominated by the plutonyl moiety in aqueous

media.1-2

The environmental dominance of the hydroxide and carbonate ligands and the toxicity

and safety issues associated with plutonium make it important to study the properties of the

plutonyl hydroxide and carbonate species. There have been several experimental and theoretical

studies of the speciation, structure and electronic properties of plutonyl carbonates.3-6

On the

other hand, hydrolysis of plutonyl complexes7-9

has been studied to a much lesser extent than that

of their uranyl counterparts.10-26

The practical effect of this is a distinct scarcity of experimental

data regarding the speciation and solution thermodynamics, structure and electronic properties of

plutonyl hydroxides in contrast to the situation for uranium. Recently, Neu et al. pointed out a

significant difference between the U(VI) and Pu(VI) hydrolysis systems.9 The plutonyl

hydrolysis system is dominated by the dimeric product, [(PuO2)2(OH)2(H2O)6]2+

, while the

uranyl system is dominated by the trimeric complex, [(UO2)3(OH)5]+, and other polynuclear

species in highly alkaline solutions. They could however not characterize the structural

Page 122: Relativistic Quantum Chemistry Applied to Actinides

100

properties of the plutonyl dimer complex using extended X-ray absorption fine structure

(EXAFS).9 In addition, they found no evidence for the formation of higher polynuclear species

such as [(PuO2)3(OH)5]+ or [(PuO2)4(OH)7]

+. It should be however noted that they used very

dilute solutions, with concentrations between 10-2

and 10-5

M Pu(VI). On the other hand, Rao et

al. recently reported their potentiometric and calorimetric work on the hydrolysis of Pu(VI) in

slightly acidic solutions at variable temperatures.8 The formation of [(PuO2)3(OH)5]

+ was

invoked in attempts to fit their calorimetric and potentiometric data. Going further back, the

formation of a tetrameric complex rather than [(PuO2)3(OH)5]+ was invoked by Madic et al. in

their evaluation of the pH dependence of the Raman active plutonyl asymmetric stretching

vibrational modes of the hydrolysis products.7

The seeming disparity in the final conclusions of these experimental works7-9

and the

general scarcity of EXAFS structural data for plutonyl complexes represent a significant gap in

our knowledge of the similarity and differences between the hydrolysis of the plutonyl group and

that of its uranium and neptunium analogues. In addition, if one adheres to the conclusions of

Neu et al.9 regarding the absence or negligible concentration of the trimeric species, there is at

the current moment only speculative insights for why this is the case as well as for why in

contrast the uranyl hydrolysis system is dominated by the trimeric and polynuclear complexes.

Theoretical calculations can however be used to provide information regarding the structural

properties and stabilities of the plutonyl hydroxo complexes. Kohn-Sham density functional

theory (DFT) is a powerful method that is rigorously only suited for single-reference systems. A

large number of open-shell actinide systems, amongst which are the Pu(VI) complexes, are

however not entirely single-reference in nature. The ground electronic states in these complexes

often contain several determinants each corresponding to particular occupation of the actinide 5f

Page 123: Relativistic Quantum Chemistry Applied to Actinides

101

orbitals. As such, the use of single-reference theoretical approaches in computing the properties

of these complexes is not fully rigorous. On the other hand, the use of DFT calculations in

predicting the structural features and vibrational frequencies of plutonyl complexes, as well of

other open-shell actinide systems, has been reported by various workers.3, 5, 27-34

The non-

bonding natures of the unpaired 5f electrons of the open-shell actinide systems ensures that there

are, to a first approximation, no differences in the bonding schemes of the different electronic

states corresponding to the different occupations of the 5f orbitals.

In this work, we have examined the structural and electronic properties of the

monomeric, dimeric and trimeric plutonyl hydroxides using scalar-relativistic DFT calculations.

The calculated structural properties of hydrolyzed plutonyl complexes are compared to available

EXAFS data35

while the calculated vibrational frequencies are compared to the Raman

spectroscopic measurements of Madic et al.7 The aim is to use theoretical calculations to fully

identify which species are present in the plutonyl hydrolysis pH speciation diagram. In addition,

the reaction energies for forming the monomeric, dimeric and polynuclear plutonyl hydroxides

are compared to those of their neptunyl and uranyl counterparts in a bid to understand the

origin(s) of the differences in the hydrolysis systems of these actinides.

Computational Details.

All DFT calculations in this work were carried out with the ADF36-37

, Gaussian 0338

and

Priroda39

codes. Calculations employing relativistic effective core potentials (RECPs) to describe

the plutonium atoms were carried out with Gaussian 03 while those employing all-electron (AE)

basis sets were carried out with the ADF and Priroda codes. In cases with difficult convergence

attributes, the plutonyl-hydroxo cores, [[PuO2(OH)n]2-n

]x, were first optimized before step-wise

Page 124: Relativistic Quantum Chemistry Applied to Actinides

102

addition of each aquo ligand. After the gas-phase geometry optimizations, vibrational

frequencies were calculated using the harmonic approximation in order to confirm the minima

nature of the optimized structures on the potential energy surfaces. The properties of the triplet,

quintet and septet multiplet states are reported for the monomeric, dimeric and trimeric plutonyl

species respectively. The deviations between the calculated and expected S(S+1) values for these

multiplet states were in all cases less than 3.5%. Attempts to examine other multiplet states, such

as the quintet state for the trimeric complexes and the antiferromagnetic singlet states for the

dimeric complexes were abandoned as a result of massive spin contaminations (greater than 10%

deviation from the exact S(S+1) value). Reaction energies were calculated as ΔE by employing

the total electronic energies of the reactants and products in an aqueous medium. In Priroda and

Gaussian 03, ultra-fine grids were used in the numerical integration of the exchange-correlation

portion of the density functionals. The grid size refers to the use of a grid-based numerical

integration step in the evaluation of the exchange-correlation energies. An integration parameter

of 6.0 was employed for the calculations carried out using the ADF code. To examine trends and

allow for comparison, a similar approach was employed for the uranium and neptunium

analogues of the plutonyl complexes studied in this work. The high-spin electronic states of the

neptunyl complexes (triplet and quartet states for the dimer and trimers respectively) were also

used.

A scalar-relativistic approximation to the full Dirac equation was used in the AE

calculations carried out with the Priroda code. In this case, all the spin-orbit terms were separated

out and neglected.40

This approach, when used with the PBE functional has been shown to lead

to accurate structural parameters, redox potentials and reaction energies for actinide

complexes.34, 41-42

A basis of double-ζ quality (cc-pVDZ) was used for all the elements for the

Page 125: Relativistic Quantum Chemistry Applied to Actinides

103

large component with the corresponding kinetically-balanced basis sets for the small

component.39

The combination of the PBE functional and this all-electron basis set is labeled as

PBE/B1. Mayer bond orders43

were calculated after the geometry optimizations. All the

calculations carried out at this level were performed in the gas-phase.

The scalar relativistic ZORA approach44-46

was employed with triple-ζ polarized (TZP)

all electron basis sets in the calculations carried out with the ADF code. No core atomic orbitals

were frozen in these calculations. The BP86 functional was used in optimizing the geometries of

all the uranyl, neptunyl and plutonyl molecules in both the gaseous and aqueous phases. This

level is labeled the BP86/B2 level. In ADF, the aqueous phase calculations were carried out

using the conductor-like screening solvation model, COSMO.47

The radii of the U, Np, Pu, O

and H atoms were taken as 2.34, 2.34, 2.34, 1.72 and 1.30 respectively. The variations of the

calculated reaction energies were found to be generally negligible for actinide solvation radii

between 2.24 and 2.44 Å.

In the RECPs calculations, the Stuttgart small-core scalar-relativistic pseudopotential was

used to describe the plutonium atoms.48-49

The pseudopotential was used to represent 60 core

electrons in plutonium while the remaining 34 electrons were represented by the associated

valence basis set. All g-type functions were removed from the valence basis set to increase

computational efficiency. The 6-31+G* basis was used to describe all the other atoms in the

molecules, except for hydrogen atoms for which the 6-31G basis was employed. This

combination of the RECP for the plutonium atoms and the 6-31+G*/6-31G bases for the non-

actinide atoms is labeled as B3. The B3LYP functional was employed in the RECP

computations. The effects of a solvent environment on the calculated reaction energies were

evaluated with single-point calculations on the geometries optimized at B3LYP/B3 level in the

Page 126: Relativistic Quantum Chemistry Applied to Actinides

104

gas-phase while employing the polarizable continuum solvation (PCM) model.50

The default

atomic radii of the united atom topological model (UA0) in Gaussian 03 were used in the

calculations. To examine trends and allow for comparison, a similar approach was employed for

the uranium and neptunium analogues of the plutonyl complexes studied in this work.

Results and Discussion

Structural Properties. The optimized structures of the monomeric plutonyl aquo-hydroxide

complexes obtained at the BP86/B2 level in the aqueous phase are shown in Figures 4.1 and 4.2.

The structural parameters and plutonyl stretching vibrational frequencies obtained for these

molecules are collected in Tables 4.1-4.4. At this level, the Pu-Oyl bonds were calculated to be

about 1.76 Å long in the pentaaquo plutonyl complex, [PuO2(H2O)5]2+

, slightly longer than the

experimental value of 1.75 Å.35

The calculated Pu-OH2 bond lengths for the equatorial ligands in

this complex are also a good match for the experimental value of 2.41 Å. The experimental Pu-

Oyl and Pu-OH2 bond distances were determined from EXAFS spectroscopy in solution.35

From

a methodological perspective, the calculated Pu-Oyl bond lengths obtained in the gas phase are

slightly shorter than those obtained in aqueous solution. The use of the hybrid functional leads to

even shorter Pu-Oyl bonds, Table 4.1.

Examination of the optimized structures of [PuO2(H2O)(OH)4]2-

, [PuO2(H2O)2(OH)3]-

and [PuO2(H2O)3(OH)2], Figure 4.1, shows that an aquo ligand in these complexes is found in

the second solvation sphere, about 3.68-4.00 Å from the plutonium atoms. Energetically, the

complexes with an aquo ligand in the second coordination sphere are generally favored in the gas

phase mostly due to the hydrogen bonding with the equatorial hydroxo ligands. This was found

to be the case at the BP86/B2 and B3LYP/B3 levels. In the aqueous phase calculations, there is a

Page 127: Relativistic Quantum Chemistry Applied to Actinides

105

dramatic reduction in the calculated binding energy of the water ligand in the 2nd

coordination

sphere. The removal of the second sphere aquo ligand from [PuO2(H2O)(OH)4]2-

,

[PuO2(H2O)2(OH)3]- and [PuO2(H2O)3(OH)2] were actually calculated to be exothermic at the

BP86/B3 level while employing the COSMO solvation model. As a result, [PuO2(OH)4]2-

,

[PuO2(H2O)(OH)3]- and [PuO2(H2O)2(OH)2], Figure 4.2, are most likely the preferred species in

solution. In addition, the calculated energies obtained in the gaseous and aqueous phases also

show that [PuO2(OH)4]2-

is preferred to its pentahydroxo counterpart, [PuO2(OH)5]3-

. It should be

noted that this does not imply the absence of a second coordination sphere. Rather, more (than

one) explicit water molecules are required for a stable second sphere.

Figure 4.1: Optimized structures of the aquo-hydroxo [PuO2(H2O)5-n(OH)n]2-n

complexes

obtained at the BP86/B2 level with the COSMO solvation model.

Page 128: Relativistic Quantum Chemistry Applied to Actinides

106

Figure 4.2: Optimized structures of the aquo-hydroxo [PuO2(H2O)4-n(OH)n]2-n

complexes

obtained at the BP86/B2 level with the COSMO solvation model.

Table 4.1: Calculated structural properties of [PuO2(H2O)5]2+

and [PuO2(H2O)4(OH)]+. The bond

distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

.

[PuO2(H2O)5]2+

[PuO2(H2O)4(OH)]+

Bond Lengths (Å) Freq. Bond Lengths (Å) Freq

Pu-Oyl Pu-OH2 Pu-Oyl Pu-OH Pu-OH2

PBE/B1 1.75 2.46 969/853 1.77 2.11 2.53-2.56 922, 814

BP86/B2

Gaseous 1.74 2.46 976/856 1.77 2.13 2.53-2.57 925, 815

Aqueous 1.76 2.41 925/827 1.78 2.14 2.50-2.53 875, 787

B3LYP/B3

Gaseous 1.72 2.47 1032/911 1.74 2.11 2.53-2.57 979, 864

Expt. 1.75 2.41 962/833 817

a EXAFS measurements in aqueous solution; Ref35

Page 129: Relativistic Quantum Chemistry Applied to Actinides

107

There is a sequential increase in the Pu-Oyl bond lengths as an aquo ligand is replaced by

a hydroxo group. This is the case at the PBE/B1, BP86/B2 and B3LYP/B3 levels of theory in

both the gaseous and aqueous phases, Tables 4.1-4.4. For example, at the BP86/B2 level in

solution, the Pu-Oyl bonds become progressively weaker by about 0.01-0.03 Å as more hydroxo

ligands are coordinated in the equatorial region. This increase in bond lengths is reflected in the

decrease in the calculated Pu-Oyl Mayer bond orders, Table 4.5. The bond orders decrease from

2.38 in the pentaaquo complex to 2.25 in the pentahydroxo complex. Although the gradual

decrease in Pu-Oyl bond strength as indicated by the bond order is small, it correlates very well

with the increasing bond lengths. This phenomenon is not unique to the plutonyl hydroxides and

similar effects have been observed in uranyl fluorides51

and uranyl hydroxides, to name a few.52

The magnitude of the Pu-Oyl bond orders indicate that the Pu-O bonds in the plutonyl moiety

have significant triple bond character and the weakening of these bonds down the plutonyl aquo-

hydroxo series could at first glance be attributed to decreasing π-bonding interactions between

the plutonium and axial oxo atoms. The Pu-OH bonds also become progressively longer down

the series with increasing number of hydroxide groups. The Pu-OH bond length increases from

2.14 Å in [PuO2(H2O)4OH]+ to 2.28Å in [PuO2(OH)4]

2- and 2.26-2.53 Å in [PuO2(OH)5]

3- at the

BP86/B2 level in the COSMO calculations. This correlates with the decreasing Pu-OH bond

orders down the series. This trend is present in both the [PuO2(H2O)4-nOH)n.H2O]2-n

and

[PuO2(H2O)4-nOH)n]2-n

series of complexes, Tables 4.5. For the terminal hydroxo complex,

[PuO2(OH)4]2-

, the calculated Pu-Oyl and Pu-OH bond lengths obtained at the BP86/B2 level are

within 0.05 Å of the EXAFS data.35

The aqueous environment as modeled by the COSMO solvation model generally tends to

alter the Pu-Oyl and Pu-OH bonds by less than 0.02 Å. In contrast the major structural effect of

Page 130: Relativistic Quantum Chemistry Applied to Actinides

108

Table 4.2: Calculated structural properties of [PuO2(H2O)3(OH)2] and [PuO2(H2O)2(OH)3]-. The

bond distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

.

[PuO2(H2O)3(OH)2] Freq. [PuO2(H2O)2(OH)3]- Freq

Pu-Oyl Pu-OH Pu-OH2 Pu-Oyl Pu-OH Pu-OH2

PBE/B1 1.79 2.18, 2.20 2.47,2.58

3.68

900/795 1.81 2.20,2.20

2.29

2.67

3.80

863/758

BP86/B2

Gaseous 1.79 2.18, 2.20 2.49,2.59

3.73

894/785 1.81 2.22,2.22

2.29

2.70

3.91

854/749

Aqueous 1.80 2.16, 2.22 2.46,2.48

3.77

1.82 2.19,2.23

2.25

2.57

3.83

B3LYP/B3

Gaseous 1.75 2.17, 2.20 2.52,2.59

3.71

951/838 1.77 2.21,2.22

2.30

2.72

3.84

913/799

the solvent environment is shortening of the Pu-OH2 bonds. We give two examples of this

phenomenon here. For [PuO2(H2O)5]2+

, the Pu-OH2 bonds were calculated to be shorter by 0.05

Å after the introduction of solvent effects at the BP86/B2 level, Table 4.1. The gas phase

structure of [PuO2(H2O)(OH)3]- is such that the lone aquo ligand interacts with one of the

hydroxo groups. This leads to a longer Pu-OH2 bond as well as one unusually long Pu-OH bond

of about 2.32 Å. Optimization in solution removes this effect, Figure 4.2, resulting in Pu-OH2

and Pu-OH bonds that are 2.58 Å and 2.21 Å long respectively, Table 4.3.

Page 131: Relativistic Quantum Chemistry Applied to Actinides

109

Table 4.3: Calculated structural properties of [PuO2(H2O)2(OH)2] and [PuO2(H2O)(OH)3]-. The

bond distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

.

[PuO2(H2O)2(OH)2] [PuO2(H2O)(OH)3]-

Bond Lengths (Å) Freq. Bond Lengths (Å) Freq.

Pu-Oyl Pu-OH Pu-OH2 Pu-Oyl Pu-OH Pu-OH2

PBE/B1 1.80 2.14 2.54 884/782 1.81 2.18,2.31 2.70 853/ 749

BP86/B2

Gaseous 1.80 2.15 2.55 875/772 1.81 2.20,2.32 2.73 842/736

Aqueous 1.81 2.15 2.48 826/748 1.82 2.21 2.58 793/ 716

B3LYP/B3

Gaseous 1.76 2.16 2.55 931/820 1.78 2.19,2.30 2.75 899/789

The calculated structural properties of the dimeric species, [(PuO2)2(OH)2(H2O)6]2+

and

[(PuO2)2(OH)4(H2O)4] are presented in Table 4.6. For these complexes, the structures of various

conformers arising from the different orientations of the terminal hydrogen atoms in the pendant

and bridging hydroxo groups were optimized and only the most stable structures are shown in

Figure 4.3. We can employ a simple approach to describing the structures of the dimeric

complexes. These complexes are μ2-dihydroxo complexes as the two plutonyl moieties are

bridged by two hydroxo ligands. To a first approximation, the bridging hydroxo groups in

[(PuO2)2(OH)2(H2O)6]2+

are shared between the PuO22+

groups, such that it appears that each

PuO22+

group is bonded to approximately one hydroxo group. On the other hand, each PuO22+

group is bonded to about two hydroxo groups in the tetrahydroxo complex,

Page 132: Relativistic Quantum Chemistry Applied to Actinides

110

[(PuO2)2(OH)4(H2O)4]. Now given this simple model, the calculated Pu-Oyl bonds in these

complexes, Table 4.6, are within the ranges suggested by the trends obtained for the

mononuclear complexes, Tables 4.1 and 4.3. Likewise the calculated Pu-Oyl bond orders in the

dihydroxo and tetrahydroxo binuclear complexes are of the same magnitude as those found in

[PuO2(H2O)4(OH)]+ and [PuO2(H2O)2(OH)2] respectively, Table 4.5.

Table 4.4: Calculated structural properties of [PuO2(OH)4]2-

and [PuO2(OH)5]3-

. The bond

distances are given in Å while the vibrational frequencies (asymmetric/symmetric plutonyl

stretching modes) are given in cm-1

.

[PuO2(OH)4]2-

Bond Lengths (Å) PuO2 Str. (cm-1

) [PuO2(OH)5]3-

PuO2 Str.

Pu-Oyl Pu-OH Asym, Sym Pu-Oyl Pu-OH Asym, Sym

PBE/B1 1.85 2.27 808, 690 1.85 2.402

BP86/B2

Gaseous 1.85 2.29 788, 699 1.84 2.46 776, 656

Aqueous 1.85 2.25 756, 683 1.85 2.26-2.53

B3LYP/B3

Gaseous 1.81 2.28 837, 727 1.80 2.45 840, 723

Expt 1.80 2.30

Page 133: Relativistic Quantum Chemistry Applied to Actinides

111

Table 4.5: Calculated Mayer bond orders of the plutonyl aquo-hydroxo complexes obtained at

the PBE/B1 level in the gaseous phase.

Pu-Oyl Pu-OHpendant Pu-OHbridging Pu-OH2

Monomers

[PuO2(H2O)5]2+

2.38 0.47

[PuO2(H2O)4(OH)]+ 2.36 1.40 0.36-0.38

[PuO2(H2O)3(OH)2] 2.34 1.04/1.15 0.39/0.50

[PuO2(H2O)2(OH)2] 2.34 1.30 0.36

[PuO2(H2O)2(OH)3]- 2.31 0.93/1.09/1.12 0.08/0.38

[PuO2(H2O)(OH)3]- 2.31 0.89-1.21 0.37

[PuO2(OH)4]2-

2.28 1.10

[PuO2(OH)5]3-

2.25 1.04

Dimers

[(PuO2)2(OH)2(H2O)6]2+

2.36 0.68 0.40

[(PuO2)2(OH)4(H2O)4] 2.34 0.90 0.66 0.40

Trimers

[(PuO2)3(H2O)6(O)(OH)3]+ 2.33-2.35 0.85 for μ-oxo 0.70 0.36

Going on to the other bonds, the Pu-OH2 bonds are slightly longer in

[(PuO2)2(OH)4(H2O)4], compared to [(PuO2)2(OH)2(H2O)6]2+

, due to the interactions between the

aquo ligands and the pendant hydroxo groups in this complex. The bonds between the plutonium

atoms and the bridging hydroxo groups were calculated to be about 2.33-2.37 Å long. At the

Page 134: Relativistic Quantum Chemistry Applied to Actinides

112

PBE/B1 level, these bonds were calculated as having bond orders of 0.65-0.68, much smaller

than the values of 0.89-1.40 obtained for the Pu-OH bonds in the monomer complexes or the

value of 0.90 obtained for the pendant Pu-OH bonds in the binuclear tetrahydroxo complex,

Table 4.5. The implication is a significant reduction in the π-donor character of the bridging OH

groups as they each now form two σ-type bonds to the two plutonium atoms. The Pu-Pu

distances in [(PuO2)2(OH)2(H2O)6]2+

and [(PuO2)2(OH)4(H2O)4] were calculated as 3.78 and 3.79

Å respectively at the BP86/B2 level in the COSMO calculations. As the covalent radius of

plutonium is 1.87 Å, implying a Pu-Pu internuclear distance of about 3.74 Å, we can conclude

that there is little covalent interaction between the actinide centers in these complexes. The

calculated bond orders associated with the Pu-Pu distances are less than 0.2 in both dimer

complexes. Tsushima et al. previously carried out EXAFS experiments on the analogous uranyl

series.53

They obtained a U-U distance of 3.88 Å for [(UO2)2(OH)2(H2O)6]2+

which was in very

good agreement with the results of the DFT calculations they had also carried out. We obtained a

U-U distance of 3.87 Å for aqueous [(UO2)2(OH)2(H2O)6]2+

at the BP86/B2 level, within 0.01 Å

of the experimental and previous theoretical work. We note the contraction of the actinide-

actinide distances by about 0.09 Å on going from the uranyl dimer complex to its plutonyl

counterpart, consistent with the actinide contraction. The covalent radius of plutonium is about

0.1 Å smaller than that of uranium.

For the trimeric complex, It has been previously noted that [(PuO2)3(H2O)6(O)(OH)3]+

and [(PuO2)3(H2O)6(OH)5]+ can‟t be distinguished from the often used acid-base titration

experiments.53

In their work on the uranyl analogue, Tsushima et al. obtained average U-U

distances of 3.81-3.82 Å from their EXAFS experiments.53

They concluded from these distances

that there is a preference for [(UO2)3(H2O)6(O)(OH)3]+ rather than [(UO2)3(H2O)6(OH)5]

+. At the

Page 135: Relativistic Quantum Chemistry Applied to Actinides

113

Table 4.6: Calculated structural properties of the plutonyl dimer complexes. The bond lengths are in Å and the calculated IR

intensities (km/mol) are given in parenthesis.

[(PuO2)2(OH)2(H2O)6]2+

[(PuO2)2(OH)4(H2O)4]

PBE/B1 BP86/B2 B3LYP/B3 PBE/B1 BP86/B2 B3LYP/B3

Gaseous Gaseous Aqueous Gaseous Gaseous Gaseous Aqueous Gaseous

Pu-Oyl 1.76 1.76 1.77 1.73 1.78 1.78 1.80 1.74-1.75

Pu-OHbridging 2.32 2.32 2.32 2.33 2.33-2.35 2.34-2.35 2.35-2.36 2.37, 2.34

Pu-OHpendant 2.30 2.30 2.20 2.26-2.32

Pu-OH2 2.51-2.53 2.52-2.55 2.43-2.52 2.53-2.55 2.58 2.61 2.59-2.60 2.58-2.59

Pu-Pu 3.82 3.80 3.78 3.84 3.71 3.74 3.79 3.75

PuO2 Str. Asym. 944,938 947, 939 1004(446),996(0) 917,908 912, 902 971(504),959(49)

PuO2 Str. Sym. 831,825 829, 823 884(0),879(33) 804,793 804, 789 852(2),845(4)

Expt. 826 a 805,793

a

Bridging OH 796,784 780, 778 844(184),838(0) 734,731 754, 748 783(205),780(13)

a Raman spectroscopic measurements; Ref7

Page 136: Relativistic Quantum Chemistry Applied to Actinides

114

Figure 4.3: Optimized structures of the bis-plutonyl aquo tetrahydroxo and dihydroxo

complexes obtained at the B3LYP/B3 level in the gas phase.

BP86/B2 level and with the COSMO solvation model, we obtained U-U distances of 3.82 Å for a

C3V structure for [(UO2)3(H2O)6(O)(OH)3]+. The agreement between the EXAFS data of

Tsushima et al.53

and the result from the BP86/B2 model increases our confidence in the use of

the C3V structure in describing the plutonyl counterpart. The structural parameters of the

optimized C3V µ3-oxo structure of [(PuO2)3(H2O)6(O)(OH)3]+ are presented in Table 4.7. This

structure is shown in Figure 4.4. Similar to the case in the [(PuO2)2(OH)4(H2O)4] dimer complex,

the Pu-Oyl bonds in this structure were calculated to be about 1.74 Å long at the B3LYP/B3 level

in the gas phase, Table 4.7. The usual strengthening of the Pu-OH2 bonds after optimization with

an implicit solvation model is also observed for this complex at the BP86/B2 level. The Pu-Pu

distances in this complex are 3.74 Å long, about 0.08 Å shorter than the U-U distances in

[(UO2)3(H2O)6(O)(OH)3]+.

Page 137: Relativistic Quantum Chemistry Applied to Actinides

115

Table 4.7: Calculated structural properties of the μ3-oxo motifs of the trimeric complex,

[(PuO2)3(H2O)6(O)(OH)3]+. The bond lengths are given in Å, bond angles in degrees while

vibrational frequencies are presented in cm-1

.

PBE/B1 BP86/B2 B3LYP/B3

Gaseous Gaseous Aqueous Gaseous

Pu-Oyl 1.78 1.78 1.79 1.74

Pu-OHbridging 2.42 2.41 2.40 2.39

Pu-μ3O 2.18 2.17 2.17 2.19

Pu-OH2 2.58 2.60 2.56 2.54

Pu-Pu 3.71 3.73 3.74 3.72

PuO2 Str. Asym. 913, 901, 903 917, 902, 902

PuO2 Str. Sym. 803, 784, 781 807,

Bridging OH 721, 715, 709 784, 748, 737

Figure 4.4: Structures of [(PuO2)3(H2O)6(O)(OH)3]+ obtained at the BP86/B2/COSMO level.

The C3V structure is shown on the left while the structure with a bridging aquo group is shown on

the right.

Page 138: Relativistic Quantum Chemistry Applied to Actinides

116

Regarding the structure of the plutonyl trimer complex there are two points of note.

Firstly, it remains to be seen whether the highly symmetrical C3V structure is the global

minimum structure on the potential energy surface. Tsushima et al. obtained another μ3-oxo

structure that was about 9.8 kcal higher in energy than the lowest energy structure they obtained

for [(UO2)3(H2O)6(O)(OH)3]+.53

This structure has two bridging hydroxo groups, a bridging aquo

group and a pendant hydroxo group. The plutonyl analogue is shown in Figure 4.4. Secondly, it

still remains to be seen whether the [(AnO2)3(O)(OH)3]+ core is preferred to the [(AnO2)3(OH)5]

+

core for plutonium as it was in the case of uranium.53

The significance of these two points will be

discussed later.

Vibrational Frequencies.

The calculated plutonyl stretching vibrational frequencies of the monomeric plutonyl

complexes are also collected in Tables 4.1-4.4. As the vibrational frequencies were calculated

without correction for anharmonic effects, this implies that deviations from available

experimental values cannot be fully avoided. The calculated structural features in Tables 4.1-4.4

show that the hybrid B3LYP functional generally over-binds the actinyl An-Oyl, bonds, relative

to the GGA functionals. The shorter An-Oyl bonds obtained with the B3LYP functional result in

vibrational frequencies that are much higher than the experimental values. This agrees well with

previous reports by other workers.33-34, 41-42, 51

On the other hand, the solvent environment tends

to reduce the calculated plutonyl stretching frequencies by about 20-40 cm-1

at the BP86/B1 and

B3LYP/B3 levels. This is also in accordance with literature experience.33, 51

Given that the PBE

functional, like most GGA functionals tends to slightly under-bind actinyl bonds in the gas-

phase, there appears to be a serendipitous coincidence between the actinyl stretching vibrational

frequencies obtained with the PBE functional in the gas-phase and the experimental values

Page 139: Relativistic Quantum Chemistry Applied to Actinides

117

obtained in the aqueous phase. The under-binding error of the GGA functional corrects for the

non-inclusion of solvation effects. Taking [PuO2(H2O)5]2+

as an example, the gas-phase PBE/B1

values for the plutonyl symmetric and asymmetric stretching modes are 853 and 969 cm-1

, Table

4.1, in good agreement with the solvent-phase experimental values of 833 and 962 cm-1

for the

symmetric and asymmetric Oyl-Pu-Oyl stretching modes respectively. The plutonyl stretching

vibrational frequencies obtained in the gas-phase with the BP86/B2 approach are also close to

the experimental value but are often lower than those obtained at the PBE/B1 level for the other

monomeric plutonyl aquo-hydroxo complexes, Tables 4.1-4.4.

To further test the usefulness of the vibrational frequencies obtained at the PBE/B1 level

for estimating experimental frequencies obtained in solution, we have calculated the vibrational

frequencies of all the hydrolyzed uranyl aquo complexes. There are available IR and Raman

experimental data for the vibrational frequencies of the uranyl hydrolysis products.22, 54

The

calculated uranyl stretching vibrational frequencies of [UO2(H2O)5]2+

, [(UO2)2(OH)2(H2O)6]2+

and [(UO2)3(H2O)6(O)(OH)3]+ obtained using the PBE/B1 approach are presented in Table 4.8.

The calculated uranyl stretching modes of [UO2(H2O)5]2+

, [(UO2)2(OH)2(H2O)2]2+

and

[(UO2)3(OH)3(O)(H2O)6]2+

are in good agreement with the experiment.22, 54

The agreement

between the calculated and experimental frequencies for these uranyl complexes, especially for

the polymeric species, gives us tremendous confidence regarding the use of the PBE/B1

approach in predicting the vibrational frequencies of the species found in the plutonyl hydrolysis

speciation diagram. The most significant deviation between the experimental and calculated

frequencies was obtained for [UO2(OH)4]2-

, Table 4.8. Overall, the use of the PBE/B1 approach

in the gas-phase is desirable as it requires little or no scaling factor due to the opposing effects of

hybrid exchange and solvation effects (as has been discussed above). It also helps that this

Page 140: Relativistic Quantum Chemistry Applied to Actinides

118

approach is based on a fully scalar-relativistic approach with large all-electron basis sets thus

avoiding the core-valence issues of RECPs.

Madic et al. have characterized the evolution of the symmetric plutonyl stretching

vibrational modes as a function of pH using Raman spectroscopy.7 The symmetric and

asymmetric plutonyl vibrational stretching modes in [PuO2(H2O)5]2+

were calculated to be at 853

and 969 cm-1

respectively at the PBE/B1 level, Table 4.1. As previously noted, the replacement

of an aquo ligand by a hydroxo ligand leads to elongation of the Pu-Oyl bonds, Tables 4.1-4.4.

This results in lower plutonyl stretching vibrational frequencies. The symmetric and asymmetric

Table 4.8: Calculated uranyl stretching vibrational frequencies (cm-1

) of uranium aquo hydroxo

complexes obtained at the PBE/B1 level.

Calculated Experimental Data22, 54

Symmetric Asymmetric Symmetric Asymmetric

[UO2(H2O)5]2+

887 973 870 962

[UO2(H2O)4]2+

894 981

[UO2(H2O)3(OH)]+ 847 921

[UO2(H2O)2(OH)2] 810 883

[UO2(H2O)(OH)3]- 777 850

[UO2(OH)4]2-

730 802 784 857

[(UO2)2(H2O)6(OH)2]2+

855, 861 944, 936 854 943

[(UO2)2(H2O)4(OH)4] 829, 838 907, 918

[(UO2)3(H2O)6(O)(OH)3]+ 822, 822, 831 898, 898, 921 836 923

Page 141: Relativistic Quantum Chemistry Applied to Actinides

119

stretching modes decrease to around 814 and 922 cm-1

respectively in the first hydrolyzed

monomer complex, [PuO2(H2O)4(OH)]+. Comparison of the speciation diagram obtained using

calculated hydrolysis constants by Neu et al.9 with that obtained by Madic et al.

7 using Raman

spectroscopic measurements, suggests that the hydrolyzed species that is first formed at around

pH 3-4 should possess a Raman active peak at 817 cm-1

. Neu et al. identified this hydrolyzed

species as [PuO2(H2O)4(OH)]+ and concluded it to be the dominant monomeric plutonyl hydroxo

complex. Our gas-phase PBE/B1 calculations predicted the Raman active symmetric Oyl-Pu-Oyl

stretching mode for this complex to be at 814 cm-1

. This stretching mode was calculated to be at

782 cm-1

in the most stable dihydroxo complex, trans-[PuO2(H2O)2(OH)2]. The trans-dihydroxo

complex is about 3.18 and 4.60 kcal/mol more stable than its cis-conformer at the BP86/B2 and

B3LYP/B3 level in solution respectively. Despite this small energy difference between the cis

and trans isomer, the experimental peak at 817 cm-1

cannot be assigned to cis-

[PuO2(H2O)2(OH)2] as it has an higher symmetric stretching vibrational frequency of 800 cm-1

. It

appears that the only possible assignment of the Raman peak at 817 cm-1

is [PuO2(H2O)4(OH)]+,

in good agreement with Neu et al.‟s conclusions.9

There is also the option that the Raman peak at 817 cm-1

could be assigned to the dimeric

species, [(PuO2)2(OH)2(H2O)6]2+

as Madic et al.7 did or even to higher polynuclear species.

There are two plutonyl groups in the dimeric complexes. For this reason, we find two sets of

asymmetric and symmetric Oyl-Pu-Oyl vibrational stretching modes described as ν2+ν2′, ν2-ν2′,

ν1+ν1′ and ν1-ν1′ with respect to the coupling of the individual plutonyl groups. For this dimeric

dihydroxo complex, the ν2+ν2′ stretching mode corresponds to the asymmetric stretching of the

Pu-Oyl bonds on both plutonyl groups in the same direction. This mode is IR active while its

counterpart, the ν2-ν2′ mode is Raman active. For the symmetric stretching modes, the ν1+ν1′

Page 142: Relativistic Quantum Chemistry Applied to Actinides

120

mode is Raman active, while the ν1-ν1′ mode is IR active. The ν2+ν2′, ν2-ν2′, ν1+ν1′ and ν1-ν1′

modes were calculated as 944, 938, 831 and 825 cm-1

respectively at the PBE/B1 level, Table

4.6. This suggests that the experimental Raman peak observed at 817 cm-1

can still be assigned to

[PuO2(H2O)3(OH)]+ and not [(PuO2)2(OH)2(H2O)6]

2+. However, the peak observed at 826 cm

-1 in

the Raman spectra7 can now be assigned to the ν1+ν1′ mode of [(PuO2)2(OH)2(H2O)6]

2+. This

matches the speciation diagram of Neu et al.9 but contravenes Madic et al.‟s

7 assignment of the

peak at 817 cm-1

to [(PuO2)2(OH)2(H2O)6]2+

. The suggestion is that even at the high Pu(VI)

concentrations that their work was carried out in, the monomeric complex could still have been

observed by Madic et al.7

Madic et al. also observed other peaks at 805 cm-1

and 794 cm-1

at slightly higher pH

values in their experimental measurements.7 Examination of the calculated vibrational

frequencies in Tables 4.6 reveals that we can assign these peaks to the ν1+ν1′ and ν1-ν1′ modes of

the tetrahydroxo dimer complex, [(PuO2)2(OH)4(H2O)4]. The Raman active O-H wagging modes

of the bridging hydroxo groups were calculated at 784 cm-1

in [(PuO2)2(OH)2(H2O)6]2+

and at

around 734 cm-1

in [(PuO2)2(OH)4(H2O)4], Table 4.6. These are most likely too low to warrant

assignment to the experimental peaks at 805 cm-1

and 794 cm-1

. From experience, we note that

the calculated vibrational frequencies obtained at the PBE/B1 level in nearly all cases either

matches or over-estimates the experimental value.33-34, 41-42, 51

There are three sets of plutonyl asymmetric and symmetric vibrational frequencies in

[(PuO2)3(H2O)6(O)(OH)3]+. These are labeled as ν2+ν2′+ν2′′, ν2-ν2′, ν2-ν2′′, and ν1+ν1′- ν1′′, ν1- ν1′

and ν1-ν1′′, reflecting the direction of the individual Pu-Oyl stretches. The symmetric vibrational

modes were calculated to be at 855, 840 and 813 cm-1

at the PBE/B1 level, Table 4.7. All these

modes were found to be IR active suggesting that this complex was not the one observed in the

Page 143: Relativistic Quantum Chemistry Applied to Actinides

121

Raman spectra of Madic et al.7 The mode calculated at 840 cm

-1 has the lowest IR intensity and

assignment to the Raman experimental peak at 826 cm-1

is tentatively possible. In addition, the

µ-hydroxo wagging vibrational frequencies are all infrared (IR) active and between 742-797 cm-

1. For this reason, they most likely cannot account for the experimental Raman

7 peaks at 794 and

805 cm-1

.

Energetics.

The reaction energies for the formation of the plutonyl dimer and trimer complexes in

solution are presented in Table 4.9. The difference in electronic energies of the reactants and

products, ΔEreaction, were employed due to the difficulty of calculating the solvent phase thermal

free energy corrections for the open-shell species at the BP86/B2 level. The calculations at the

BP86/B2 level are expected to be more accurate as they use a triple-ζ basis set, thereby reducing

basis set superposition errors, and include the effect of the aqueous environment on the structure

of the relevant complexes. In contrast the B3LYP/B3 results were obtained from single-point

calculations on the gas-phase geometries. The calculated energies presented in Table 4.9 are for

the solvent phase reactions and as a result do not include corrections for basis set superposition

errors (BSSE). Counterpoise calculations were carried out to estimate these errors for the

dimerization of the [AnO2(H2O)2(OH)2] species in the gas phase. They amount to around 1.1-1.8

kcal/mol at the B3LYP/B3 level and are much lower (0.4-0.6 kcal/mol) at the BP86/B2 level.

This is understandable given the triple-ζ nature of the basis sets employed in the BP86/B2

calculations.

Formation of the [(AnO2)2(OH)2(H2O)6]2+

complexes by dimerization of the monomeric

[AnO2(H2O)4(OH)]+ complexes was calculated to be more endothermic than the agglomeration

Page 144: Relativistic Quantum Chemistry Applied to Actinides

122

of the pentaaquo complex with [AnO2(H2O)4(OH)2] species. This is most likely an electrostatic

effect. The production of the [(AnO2)2(OH)4(H2O)4] complexes by dimerization of the

[AnO2(H2O)2(OH)2] species were calculated to be less exothermic than the formation of their

counterpart binuclear dihydroxo complexes. Examination of Table 4.9 shows that the formation

of the trimer complexes, is more exothermic for the uranyl and neptunyl complexes by about 6

kcal/mol in comparison to the plutonyl complex. Assuming that thermochemical factors are

dominant over kinetic factors, one can say from the calculated energies that the significance of

the [(AnO2)3(H2O)6(OH)3(O)]+ species follows the trend U ≈ Np > Pu.

It is however not fully clear that the reactions listed in Table 4.9 represent the dominant

mechanisms for the formation of these bimetallic and trimetallic complexes. For this reason, the

direct hydrolyses of the monomeric pentaaquo complexes were also studied and the reaction

energies are compiled in Table 4.10. The calculation of the reaction energies of the hydrolysis of

the actinyl aquo complexes (estimation of their pKa values) is rather difficult due to the issues

regarding estimating the solvation free energies of a proton.55-57

In this work we have used the

hydronium ion in the reactions presented in Table 4.10. In aqueous solution, the driving force for

the formation of the hydronium and hydroxide ions from water was calculated as 62.9 and 67.3

kcal/mol at the BP86/B2 and B3LYP/B3 levels respectively. These values correspond to pKa

values of 21.7 and 20.3 respectively in comparison to the experimental value of 15.7.

For the monomeric aquo-hydroxo complexes, the reactions become increasingly

endothermic as the degree of hydrolysis is increased. The formation of [UO2(H2O)4(OH)]+ and

[PuO2(H2O)4(OH)]+ were calculated to be endothermic in solution by about 10.1 and 9.5

kcal/mol respectively at the BP86/B2 level. These agree moderately well with the experimental

values of 7.1 and 8.0 kcal/mol respectively. The possible sources of the residual errors in the

Page 145: Relativistic Quantum Chemistry Applied to Actinides

123

Table 4.9: Calculated reaction energies (kcal/mol) for the formation of the dimer and trimer

actinyl hydroxo complexes in the aqueous phase.

calculated reaction energies are known (such as basis set errors, solvation model, reaction model,

approximate nature of Kohn-Sham DFT and problems with the accuracy of the experimental

reaction energies). For the dimer complexes, the reactions leading to the formation of the

dihydroxo species are significantly less endothermic than similar reactions for the

Reactions U Np Pu

Dimers

2AnO2(H2O)4(OH)+ → (AnO2)2(H2O)6(OH)2

2+ + 2H2O

B3LYP/B3 -4.8 -1.9 -7.5

BP86/B2 -4.1 2.4 -4.2

AnO2(H2O)2(OH)2 + AnO2(H2O)52+

→ (AnO2)2(H2O)6(OH)22+

+ H2O

B3LYP/B3 -20.2 -17.1 -29.0

BP86/B2 -17.8 -11.6 -15.3

2AnO2(H2O)2(OH)2 → (AnO2)2(H2O)4(OH)4

B3LYP/B3 -12.6 -7.6 -2.0

BP86/B2 -12.9 0.3 -5.1

Trimers

(AnO2)2(H2O)4(OH)4 + AnO2(H2O)4(OH)+ → (AnO2)3O(H2O)6(OH)3

+ +3H2O

B3LYP/B3 -8.5 -8.1 -2.0

BP86/B2 -13.8 -13.9 -7.8

Page 146: Relativistic Quantum Chemistry Applied to Actinides

124

tetrahydroxides. This would suggest that the dihydroxo species are most likely the dominant

binuclear species in solution. Overall for the dimer and monomer aquo-hydroxo species, the

reaction energies of the plutonium complexes are generally within 1.5 kcal/mol of those of their

uranium counterparts. The formation of the trimer complexes is least endothermic for

[(UO2)3(H2O)6(OH)3(O)]+ and most endothermic for [(PuO2)3(H2O)6(OH)3(O)]

+. Similar to the

case in Table 4.9, there is a difference of about 6 kcal/mol between uranyl and plutonyl systems,

Table 4.10.

Table 4.10: Calculated reaction energies (kcal/mol) obtained actinyl hydrolysis reactions at the

BP86/B2 while using the COSMO solvation model.

Reactions U Np Pu

Monomers

AnO2(H2O)52+

+ H2O → AnO2(H2O)4(OH)+ + H3O

+ 10.1 7.7 9.5

Experimental8, 26

7.1 8.0

AnO2(H2O)52+

+ H2O → AnO2(H2O)2(OH)2 + 2H3O+ 33.8 29.4 30.2

AnO2(H2O)52+

+ 2H2O → AnO2(H2O)(OH)3- + 3H3O

+ 63.8 60.1 63.0

AnO2(H2O)52+

+ 3H2O → AnO2(OH)42-

+ 4H3O+ 101.3 97.8 101.8

Dimers

2AnO2(H2O)52+

→ (AnO2)2(H2O)6(OH)22+

+ 2H3O+ 16.0 17.8 14.9

2AnO2(H2O)52+

+ 2H2O → (AnO2)2(H2O)4(OH)4 + 4H3O+ 54.6 59.1 55.3

Trimers

3AnO2(H2O)52+

→ (AnO2)3O(H2O)6(OH)3+ + 5H3O

+ 50.9 53.1 57.1

Page 147: Relativistic Quantum Chemistry Applied to Actinides

125

The consistency between the approximately 6 kcal/mol difference between the uranyl and

plutonyl trimer systems raises some questions. Firstly, how significant is 6 kcal/mol or an

equilibrium constant of 1, via e-ΔG/RT

? To recast this question, one wonders if 6 kcal/mol is

enough to explain the low/negligible concentration of the plutonyl trimer in contrast to the

significant mole fraction of the trimer in the uranyl hydrolysis system. This discrepancy in the

percentage concentration of the trimeric uranyl and plutonyl complexes has been observed

experimentally by Neu et al.9 and Rao et al.

8, 26 Fortunately, Rao et al., as part of their long line

of work on reaction energetics of actinide complexes58-62

, have reported experimental energy

data for the uranyl and plutonyl hydrolysis systems. The difference between the experimental

hydrolysis reaction energies leading to the uranyl trimer (89.8 kJ/mol) and the plutonyl trimer

(116.3 kJ/mol) is 26.4 kJ/mol (6.3 kcal/mol).8, 26

The theoretically calculated difference in

reaction energies between these actinide trimer systems, ΔΔEreaction, Tables 4.9 and 4.10,

therefore coincides with the experimental work of Rao et al.8, 26

This coincidence, most likely a

result of error cancellation in the theoretical and experimental data sets, suggests that 6 kcal/mol

is enough to explain the disparity in the experimental mole fractions of the uranyl and plutonyl

trimer complexes. As a point of note, the gas-phase reaction energies (both ΔEreaction and

ΔGreaction) favor the formation of the uranyl trimer complex by up to 11.7 kcal/mol.

Secondly, it is important that the origin of this difference be explained. One possible

suggestion would be that the septet electronic state does not represent the most stable multiplicity

for the plutonyl trimer complex. Single point calculations with total multiplicities of 3 and 5

were carried out at the septet electronic state geometry at the PBE/B1, BP86/B2 and B3LYP/B3

levels in the gaseous and aqueous phases. All these states were found to be higher in energy than

the septet state. Unfortunately, spin contaminations from the low-lying septet state make most of

Page 148: Relativistic Quantum Chemistry Applied to Actinides

126

the calculated energies for these states unreliable. The lowest spin contaminations were obtained

at the PBE/B1 level. At this level, full optimization for total multiplicities of 3 and 5 for the

plutonyl trimer complex shows that they are about 10.3 and 13.3 kcal/mol higher in energy than

the septet state respectively. We can conclude with sufficient confidence that the presence of

other electronic states lower in energy than the septet state is most likely not the origin of the

phenomenon of differing trimer mole-fractions between the uranyl and plutonyl series.

The origin of this discrepancy between the hydrolysis systems has been speculated to be

due to the greater oxophilicity of U(VI) in comparison to Pu(VI) by Neu et al.9 Looking at the

structure of [(PuO2)3(H2O)6(OH)3(O)]+, Figure 4.4, this difference in oxophilicity is possibly

manifest in the differences in the amount of stability conferred by the dehydration of the

[(AnO2)3(OH)5]+ complexes to form the [(AnO2)3(O)(OH)3]

+ species. As we previously noted,

both the [(AnO2)3(OH)5]+ and [(AnO2)3(O)(OH)3]

+ species consume five protons upon titration

and as such can‟t be distinguished using potentiometric methods. We optimized the structures of

the bare [(AnO2)3(OH)5]+ complexes at the B3LYP/B3 level in the gas phase and calculated the

energies for their dehydration in solution. The most stable structures for the [(AnO2)3(OH)5]+

complexes feature a central μ3-hydroxo and a pendant hydroxo in addition to three bridging

hydroxo groups. The dehydration energies for [(UO2)3(OH)5]+, [(NpO2)3(OH)5]

+, and

[(PuO2)3(OH)5]+ to form the respective [(AnO2)3(O)(OH)3]

+ complex were calculated as -24.7, -

16.7 and -3.5 kcal/mol respectively. The μ3-oxo atom therefore stabilizes the trimeric hexagonal

core to a greater extent in the uranyl complex than it does in the plutonyl complex.

A scheme depicting the fragmentation of the trimer complex into the component dimer

and monomer units is shown in Figure 4.5. In this scheme, the structures with bridging aquo

group, labeled 5b, are involved in formation and break-up of the trimetallic core. We note that

Page 149: Relativistic Quantum Chemistry Applied to Actinides

127

the oxophilicity argument via dehydration of the [(AnO2)3(OH)5]+ complexes minimizes the

possible role of the 5b complexes. In the final analysis, the [(PuO2)3(OH)5]+ complex as well as

the 5b species should be closer in energy to the hexagonal C3V structure, 5a, than the case in the

analogous uranium system. This would imply that the formation of the trimeric core is more

endothermic in the plutonyl system. The converse (easier fragmentation of the trimeric core)

Figure 4.5: Scheme depicting the formation and decomposition of the μ3-oxo hexagonal

trimetallic core of the trimer aquo-hydroxo complexes.

Page 150: Relativistic Quantum Chemistry Applied to Actinides

128

would be true for the backward reaction. The structure of the 5b actinyl aquo-hydroxo

complexes were optimized at the BP86/B2 level in aqueous solution, Figures 4.4 and 4.5. At this

level, these structures are about 8.3, 5.6 and 0.9 kcal/mol higher in energy than the 5a structures

for the U, Np and Pu systems respectively.

It can be concluded that the higher-energy difference between the 5a and 5b structures

for the uranyl trimer complex and to some extent the significant stability conferred by the μ3-oxo

atom (after dehydration of [(UO2)3(OH)5]+) is responsible for its persistence in solution. The

reverse of these are true for the analogous plutonyl system. The 5a structure was found to be iso-

energetic (0.9 kcal/mol) with its 5b counterpart. The 5b plutonyl complex also more easily

transforms into the [(AnO2)3(OH)5]+ complex, 3.5 kcal/mol versus 24.7 kcal/mol for the uranyl

system, which can then fragment to the dimer and monomer components.

Electronic Structure.

Electronically, the valence molecular orbitals (MOs) of the plutonyl moiety are a pair of

degenerate 5f orbitals, the σ(f) orbital, the σ(d) orbital and a pair each of π(f) and π(d) orbitals,

Figure 4.6. This orbital framework is somewhat similar to that of the uranyl and neptunyl

moieties with the exception of two degenerate singly occupied 5f orbitals. A comparison of the

electronic structures of the plutonyl aquo-hydroxo complexes to those of their uranyl and

neptunyl counterparts is interesting. The calculations at the BP86/B2 level using the COSMO

model show that the An-Oyl, An-OH and An-OH2 bonds of the [AnO2(H2O)n(OH)4-n]2-n

complexes undergo a small contraction of about 0.03 Å between the uranium and plutonium

analogues. The shorter An-Oyl bonds in PuO22+

compared to UO22+

can be explained by larger 5f

Page 151: Relativistic Quantum Chemistry Applied to Actinides

129

contributions to the σ(f) and π(f) orbitals and well as lower 6p contributions to the σ(f) orbitals,

Figure 4.6.

Figure 4.6: The α-spin molecular orbitals of [PuO2]2+

. All the orbitals contained in this picture

are occupied by one electron. In order of decreasing energies, these are the two 5f orbitals (2.28

eV, 100% Pu-5f), the σ(f) orbital (0.00 eV, 61.4% 2p from each oxo atom, 14.4% 5f from Pu and

6.5% 6p from Pu), the σ(d) orbital (-0.45 eV, 38.7% 2p from each oxo atom, 12.1% 6d from Pu,

1% 7s from Pu), the two π(d) orbitals (-0.50 eV, 41.3% 2p from each oxo atom, 14.8% 6d from

Pu) and the two π(f) orbitals (-0.63 eV, 43.7% 5f from Pu, 26.9% 2p from each oxo atom).

Page 152: Relativistic Quantum Chemistry Applied to Actinides

130

For example, the 5f orbital contribution to the σ(f) orbital increases from 57.1% to 57.5% and

61.4% down the U, Np and Pu series while the 6p orbital contribution decreases from 9.7% to

6.6%. Associated with the decreasing 6p and increasing 5f contributions to the σ(f) MO is its

stabilization relative to the σ(d) orbital on going from UO22+

to PuO22+

. As a result of this

stabilization, the σ(f) orbital is about 0.45 eV higher in energy than the σ(d) orbital in plutonyl in

comparison to the case in uranyl and neptunyl where it is about 1.04 eV and 0.78 eV higher in

energy respectively, Figure 4.6. The 5f contribution to the π(f) orbitals also increase down the

series (31.2%, 37.3% and 43.7%). The resulting stabilization of the π(f) orbitals down the series

actually results in a switching of the ordering of the π(f) and π(d) orbitals in plutonyl, Figure 4.7.

Figure 4.7: The energy levels of the [AnO2]2+

optimized at the BP86/B2 level with the COSMO

model. Only the α-spin molecular orbitals (1 electron each) are shown for NpO22+

and PuO22+

.

Page 153: Relativistic Quantum Chemistry Applied to Actinides

131

The energy separation between these orbitals is 0.24 eV, 0.09 eV and -0.13 eV respectively in

UO22+

, NpO22+

and PuO22+

. The stabilization of the σ(f) and π(f) orbitals down the series is also

seen at the B3LYP/B2 level.

The coordination of hydroxo ligand(s) to the plutonyl moiety results in significant

stabilization of the σ(d) orbital as it is the lowest plutonyl MO in the [PuO2(H2O)4-n(OH)n]2-n

complexes, Figure 4.8. The σ(d) orbital in [PuO2(OH)4]2-

has some minor bonding contributions

(about 1.5% 2p each) from the equatorial ligands. The stabilization of the actinyl σ(d) orbital has

been previously observed by Kaltsoyannis et al.52

in their study of uranyl aquo-hydroxo

complexes. The σ(f) orbital at -2.28 eV is also stabilized on hydrolysis but to a much lesser

extent than its σ(d) counterpart. The bonding overlap with the four OH ligands (about 4% 2p

each) is stronger in the σ(f) orbital than in the σ(d) one, Figure 4.8. In contrast, the energies of

the π(f/d) orbitals are relatively unchanged between the bare cation and the tetrahydroxide. The

description of these orbitals have however been altered in the hydroxide complex. The Pu-6d

contributions to the plutonyl π orbitals are much lower in [PuO2(OH)4]2-

than in PuO22+

. The Pu-

6d contributions are about 2.7% in the π(f) orbitals at -3.35 eV and 4.7% in the π(f/d) orbitals at -

2.97 eV, Figure 4.8. These are in contrast to the 14.8% Pu-6d contribution to the π(d) orbitals at -

0.50 eV in PuO22+

, Figure 4.6. The Pu-OH bonds are characterized by a pair of π-type orbitals at

-2.24 eV. The orbital at -3.06 eV is of mixed σ(Pu-OH)/σ(Pu-Oyl) character due to 2p

contributions of about 13% from each oxo atom as well as 14.2% 2p contributions from each

hydroxo group. The decreased Pu-6d contributions to the PuO22+

π-orbitals (-0.50 and -0.63 eV

in Figure 4.6), increased Pu-5f contributions to the Pu-OH π orbitals (-2.24 eV in Figure 4.8) and

increased 6d contributions to the mixed σ(Pu-OH)/σ(Pu-Oyl) orbitals down the [PuO2(OH)n]2-n

series is a good explanation for the elongation of the Pu-Oyl down the series. There is essentially

Page 154: Relativistic Quantum Chemistry Applied to Actinides

132

a transfer of actinyl 6d and 5f electron densities from the Pu-Oyl to Pu-OH bonds. This

explanation is also is in line with Clark et al.‟s speculation regarding the origin of a similar

phenomenon in the analogous uranyl series.63

Figure 4.8: Selected occupied alpha spin MOs of [PuO2(OH)4]2-

. In order of decreasing energies,

these are the two 5f orbitals (0.00 eV, 88% Pu-5f; -0.14 eV, 100% Pu-5f), the π(Pu-OH) orbitals

(-2.24 eV, 29.5% 2p from each from two OH, 18% Pu-5f, 5% 2p from other OH), the σ(Pu-O

and Pu-OH mixed) orbital (-2.28 eV, 4% from each OH, 36.6% Pu-5f, 5.7% Pu-6p, 16.7% from

each oxo atom), the two π(f/d) orbitals (-2.97 eV, 71% and 4% 2p from oxo atoms, 6.0% Pu-5f

Page 155: Relativistic Quantum Chemistry Applied to Actinides

133

and 4.7% Pu-6d), the σ(Pu-O and Pu-OH mixed) orbital (-3.06 eV, 11.7% Pu-5f, 1.0% Pu-6p,

5.7% Pu-6d, 13% 2p from each oxo and 14.2% 2p from each OH), the two π(f) orbitals (-3.35

eV, 58.5% and 2% 2p from the oxo atoms, 21.1% Pu-5f, 2.7% Pu-6d, 5% 2p from each OH) and

the σ(d) orbital (-3.86 eV, 37.0% 2p from each oxo, 11.8% Pu-6d and 1.5% 2p from each

hydroxo).

As noted for the bare actinyl complexes, there is a switching of the ordering of the π(d)

and π(f) orbitals between NpO22+

and PuO22+

due to the stabilization of the σ(f) and π(f) orbitals

across the U, Np and Pu series, Figure 4.7. In [UO2(OH)4]2-

, the lower energy π orbitals at 0.48

eV above the σ(d) orbital, have about 6.0% U-6d and 9.1% U-5f contributions. On the other

hand, the higher energy π orbitals at 0.88 eV, have 9.9% U-5f and 3.2% U-6d contributions. In

the plutonyl analog, we find that the higher energy π orbitals (at -2.97 eV in Figure 4.8) contains

about 6.0% Pu-5f and 4.7% Pu-6d atomic contributions while the lower π orbitals (at -3.35 eV in

Figure 4.8) contain 21.1% Pu-5f and 2.7% Pu-6d contributions. This indicates that the more

stable π orbitals in [PuO2(OH)4]2-

are mainly of π(f) character in contrast to their mixed π(f/d)

natures in the uranyl analogue. For the neptunyl analogue, the lower π orbital is 5.1% Np-6d and

12.6% Np-5f while the less stable π orbital is 4.3% Np-6d and 9.2% Np-5f. This switching in

orbital character, especially for the π(f) and σ(f) orbitals across the U, Np and Pu series is also

present in the aquo-hydroxo species intermediate between the bare dication and the tetrahydroxo

species. The stabilization of the σ(f) orbital and the switching of the ordering of the π(f)/π(f/d)

orbitals across the U, Np and Pu series of tetrahydroxo complexes is presented in Figure 4.9.

Page 156: Relativistic Quantum Chemistry Applied to Actinides

134

We also note that for [PuO2(OH)4]2-

, the mixed σ(Pu-OH)/σ(Pu-Oyl)-type orbital at -3.06

eV in Figure 4.8 somewhat resembles the σ(f) orbital at -2.28 eV. The orbital at -3.06 eV

however has lower Pu-5f and O-2p contributions from the plutonyl entity. It has greater

contributions from the equatorial hydroxo groups. There is a slight destabilization of this orbital

down the U, Np and Pu series, Figure 4.9. Working backwards to [UO2(OH)4]2-

, we find that the

mixed σ(Pu-OH/Pu-O)-type orbital has higher contributions from the hydroxo groups and less

contributions from the plutonyl group suggesting it has greater An-OH character than the one

found in the plutonyl analogue.

Figure 4.9: The energy levels of the [AnO2(OH)4]2+

optimized at the BP86/B2 level with the

COSMO model.

Page 157: Relativistic Quantum Chemistry Applied to Actinides

135

The monomeric plutonyl complexes possess one σ(f), one σ(d), two π(f) and two π(d),

orbitals of alpha spin in an unrestricted DFT picture. It is therefore logical to expect two σ(f),

two σ(d) and eight π(f/d) orbitals in the dimeric complexes. For simplicity, we select to analyze

the electronic structure of gas-phase [(PuO2)2(OH)2]2+

, Figure 4.10. The aqueous phase

electronic structure is qualitatively similar, especially regarding the atomic composition of the

MOs. Overall, the general description of the electronic structure of this dimer complex and its

tetrahydroxo counterpart, [(PuO2)2(OH)4], are consistent at the B3LYP/B3 and BP86/B2 levels.

MO-39 and MO-40 constitute the O-H bonds of the bridging hydroxo groups. The character and

atomic contributions to these orbitals at the B3LYP/B2 level are presented in Table 4.11. They

are semi-core orbitals found at about 1.6-2.2 eV and 2.2-2.4 eV below the valence actinyl

orbitals at the BP86/B2 and B3LYP/B3 levels respectively. These orbitals are respectively

bonding and antibonding across the bridging HO-OH groups. The π(d) and π(f) orbitals are

found between 3.02 and 2.61 eV below the HOMO, Table 4.11. It appears that orbitals with

contributions from the bridging OH groups are stabilized to a greater extent than those without.

This was found to be the case in both the gaseous and aqueous phases. Several of the π orbitals

found in this complex also possess σ(Pu-OH) [MO-42, MO-44] and σ(d) [MO-48] characters.

The σ(f) and σ(d) orbitals of plutonyl groups are found at slightly higher energies between -2.53

and -2.34 eV, Table 4.11. The σ(f) orbitals have contributions of about 3-4% from Pu 7p atomic

orbitals while the σ(d) orbitals have admixtures of Pu-5f contributions of up to 9% in MO-48 and

5% in MO-52. The admixture of Pu-7p orbitals in σ(f) MOs is also seen in the monomeric

hydroxides, as so is the admixture of Pu-5f and Pu-6d orbitals in π(d), σ(d) and π(f) MOs,

Figures 4.6 and 4.8. The bridging Pu-OH bonds in [(PuO2)2(OH)2]2+

are supported by MO-53

and MO-54 which are formed by overlap of OH-2p and Pu-5f atomic contributions. There are

Page 158: Relativistic Quantum Chemistry Applied to Actinides

136

some Oyl-2p contributions to these orbitals. The remaining occupied orbitals are the four 5f1

orbitals and the two orbitals for the lone electron pairs of the hydroxo ligands. The OH lone pair

orbitals contain about 5-9% Pu 5f while MO-59, the 5f1 orbital with highest energy has

significant OH lone pair character. The first sets of virtual orbitals are empty 5f orbitals.

Figure 4.10: The 39th

to 62nd

MOs of [(PuO2)2(OH)2]2+

of alpha spin.

Page 159: Relativistic Quantum Chemistry Applied to Actinides

137

Table 4.11: Energies, atomic contributions and descriptions of the MOs of [(PuO2)2(OH)2]2+

at

the B3LYP/B3 level. The orbital energies are scaled such that the HOMO is at 0.00 eV.

MO Energy, eV Atomic Contributions Description

39 -5.19 3% 6d each Pu, 33% 2p each OH, 7% 1s each H σ(H-O-O-H)

40 -4.64 3% 5f each Pu, 2% 6d each Pu, 33% 2p each OH,

7% 1s each H

σ(O-H)

41 -3.02 10% 6d each Pu, 15% 2p each Oyl, 7% 2p each

OH

π(d)

42 -2.94 8% 5f each Pu, 5% 6d each Pu, 12% 2p each Oyl,

10% 2p each OH

π(f/d), σ(Pu-OH)

43 -2.90 10% 6d each Pu, 18% 2p each Oyl, 5% 2p each

OH

π(d)

44 -2.86 16% 5f each Pu, 12% 2p each Oyl, 8% 2p each OH π(f), σ(Pu-OH)

45 -2.75 3% 5f each Pu, 9% 6d each Pu, 19% 2p each Oyl, π(d)

46 -2.72 13% 5f each Pu, 3% 6d each Pu, 16% 2p each Oyl, π(f)

47 -2.72 7% 5f each Pu, 8% 6d each Pu, 17% 2p each Oyl, π(d/f)

48 -2.64 9% 5f each Pu, 5% 6d each Pu, 18% 2p each Oyl, π(f),σ(d)

49 -2.61 20% 5f each Pu, 14% 2p each Oyl, π(f)

50 -2.53 27% 5f each Pu, 3% 6p each Pu, 8% 2p each Oyl,

3% 2p each OH

σ(f)

51 -2.38 26% 5f each Pu, 4% 6p each Pu, 9% 2p each Oyl, σ(f)

52 -2.34 5% 5f each Pu, 2% 2p each OH, 19% 2p each Oyl,

6% 6d each Pu

σ(d)

Page 160: Relativistic Quantum Chemistry Applied to Actinides

138

53 -1.72 13% 5f each Pu, 25% 2p each OH, 4% 2p each

Oyl, 4% 6d each Pu

σ(Pu-OH)

54 -1.59 26% 5f each Pu, 14% 2p each OH, 5% 2p each Oyl σ(Pu-OH)

55 -1.15 44% 5f each Pu, 6% 2p each OH 5f

56 -1.02 48% 5f each Pu 5f

57 -0.89 40% 5f each Pu, 8% 2p each OH 5f

58 -0.55 36% 2p each OH, 9% 5f each Pu OH lone pair

59 -0.45 24% 5f each Pu, 20% 2p each OH, 2% 2p each Oyl 5f + OH lone pair

60/HOMO 0.00 38% 2p each OH, 5% 5f each Pu, 3% 2p each Oyl OH lone pair

61/LUMO 3.55 49% 5f from each Pu 5f

62 3.67 48% 5f from each Pu 5f

The electronic structures of the trimer complexes, [(PuO2)3(H2O)6(OH)3(O)]+ and

[(PuO2)3(OH)3(O)]+ are too complicated to be fully described here. 18 actinyl (6 σ and 12 π)

orbitals are expected in these complexes, with these orbitals containing atomic contributions

from each plutonyl group as well as from the bridging hydroxo and μ3-oxo groups. This is

similar to the case of the dimeric dihydroxo complex. As an example, MO-72 in

[(PuO2)3(OH)3(O)]2-

, Figure 4.11, consists of π(f) contributions from each plutonyl group. The

presence of the μ3-oxo bridge introduces not only oxo-bridge lone-pair orbitals but also results in

the modification of some orbitals of the plutonyl groups. An example of the effect of the

bridging oxo-group on the electronic structure is depicted in Figure 4.11. MO-63 at around -

15.85 eV, Table 4.12, contains π(d) type contributions from each of the plutonyl groups which

Page 161: Relativistic Quantum Chemistry Applied to Actinides

139

overlap with 2p orbitals of the bridging oxo. The participation of the atomic 2p orbitals of the μ3-

oxo atom to this particular orbital amounts to about 18%, Table 4.12.

Figure 4.11: Selected alpha spin MOs of the trimer complex, [(PuO2)3(O)(OH)3]+.

We have previously noted that subsequent to the formation of the [(AnO2)3(OH)5]+

species, dehydration to the μ3-oxo complex (5a in Figure 4.5) with three bridging hydroxo

ligands confers significant stability on the uranyl trimer complex, [(UO2)3(O)(OH)3]+, in contrast

to its plutonyl analog, [(PuO2)3(O)(OH)3]+. The uranyl complex with a bridging aquo group (5b

in Figure 4.5) is also less stable relative to the 5a complex in comparison to the plutonyl system.

It would be quite interesting to see whether the origin of this discrepancy can be found by

examining the electronic structure of the oxo-trihydroxo trimer complexes. In the

[(AnO2)3(O)(OH)3]+ complexes, MOs 63-65 are actinyl-π(f/d) orbitals with significant σ(An-

μ3O) characters. These orbitals are shown between -15.5 and -16.0 eV in Figure 4.12. The

Page 162: Relativistic Quantum Chemistry Applied to Actinides

140

contributions from the μ3-oxo to these three orbitals are highest in the U(VI) complex and lowest

in the Pu(VI) trimer, Table 4.12. The greater participation of the 2p atomic contributions of the

μ3-oxo atom to these orbitals correlates well with the greater oxophilicity of uranium centre and

the role of the central oxo atom in further stabilizing the hexagonal (UO2)3(OH)3 core. To further

emphasize this particular point, we note that the splitting between MO-63 (the π(d)/μ3O-2p

orbital) and MO-64/MO-65 (the π(f/d)/σ(Pu-μ3O) orbitals) were calculated as 0.21, 0.26 and 0.33

eV for the U, Np and Pu complexes respectively at the B3LYP/B2 level in the gas phase.

MOs 66-67 are of actinyl-σ(d) character in the trimer complexes, Table 4.12. They also

possess some σ(An-OH) characters. The 5f contributions to these orbitals however increase from

0-2% to 4-5 % and 3-8% in the U(VI), Np(VI) and Pu(VI) trimers respectively. The increase in

5f contributions down the series is also is noticeable for the actinyl-π and σ(f) orbitals, Figure

4.12. For example, in the U(VI) complex, the π(d) orbitals (MO 68-72) have negligible U-5f

contributions. In contrast, there are significant π(f)-π(d) mixing in the plutonyl complex. This

π(f)-π(d) mixing down the series results in further stabilization of the Oyl=An=Oyl bonds and is

reminiscent of the previously discussed switching of the energetic ordering of the π(f) and π(d)

orbitals in the bare dications and monomer tetrahydroxo complexes on going from U to Pu,

Figures 4.7 and 4.9.

Page 163: Relativistic Quantum Chemistry Applied to Actinides

141

Table 4.12: Energies (eV), individual atomic contributions and descriptions of the MOs of

[(PuO2)3(O)(OH)3]+ at the B3LYP/B3 level. The orbital energies are scaled such that the HOMO

is at 0.00 eV. Scaling allows for easier presentation.

MO Atomic Participations Description Energy

63 7% Pu-6d, 5-12% Oyl 2p, ~2% μ-OH, 18% μ3(O) 2p π(d)/μ3(O-2p) -15.85

64 3-6% Pu-5f, 3-5% Pu-6d, 23% μ3(O) 2p, 8-11% Oyl 2p,

4% μ-OH

π(f/d), σ(Pu-μ3O)

-15.52

65 4-8% Pu-5f, 4-5% Pu-6d, 22% μ3(O) 2p, 8-13% Oyl 2p, π(f/d), σ(Pu-μ3O) -15.51

66 4% Pu-5f, 11% Pu-6d, 13% Pu-7s, 20-26% Oyl 2p, 3% μ-

OH

σ(d) Pu′-Oyl

-15.09

67 3-8% Pu-5f, 9% Pu-6d, 12% Pu-7s, 14-20% Oyl 2p, 2%

μ-OH

σ(d) Pu′-Oyl

π(f) Pu′′-Oyl

-15.08

68 6-7% Pu-5f, 7% Pu-6d, 9% Pu-7s, 10-19% Oyl 2p π(d) -15.05

69 3-11% Pu-5f, 3-8% Pu-6d, 3-29% Oyl 2p π(f) -15.00

70 8-9% Pu-5f, 6% Pu-6d, 7-27% Oyl 2p π(f/d) -14.97

71 3-4% Pu-5f, 3-7% Pu-6d, 3-30% Oyl 2p π(d/f) -14.91

72 10-18% Pu-5f, 3% Pu-6d, 2% Pu-6p, 3-15% Oyl 2p π(f) -14.83

73 3-15% Pu-5f, 2-3% Pu-6d, 3-13% Oyl 2p, 3% μ3(O) 2p π(f)/σ(f) -14.82

74 2-14% Pu-5f, 2% Pu-6d, 7-17% Oyl 2p, 3% μ3(O) 2p π(f)/σ(Pu-(μ3-O)) -14.81

75 3-7% Pu-5f, 2-4% Pu-6d, 5-19% Oyl 2p, 3% μ3(O) 2p π(f)/σ(f) -14.79

76 7-11% Pu-5f, 2% Pu-6d, 2-21% Oyl 2p, 2% μ-OH π(f) -14.78

77 7-9% Pu-5f, 2-3% Pu-6d, 11-25% Oyl 2p, 2% μ-OH π(f) -14.71

78 4-13% Pu-5f, 3% Pu-6d, 2-38% Oyl 2p, 2% μ-OH σ(f) -14.69

Page 164: Relativistic Quantum Chemistry Applied to Actinides

142

79 10-24% Pu-5f, 3% Pu-6d, 2-21% Oyl 2p, σ(f) -14.44

80 2-20% Pu-5f, 2% Pu-6d, 2-3% Pu-6p, 6-17% Oyl 2p σ(f) -14.44

Figure 4.12: Abbreviated energy level diagram of the [(AnO2)3(O)(OH)3]+ complexes.

Conclusions

We have presented a comprehensive examination of the structure and electronic

properties of monomeric, dimeric and trimeric plutonyl aquo-hydroxo complexes using scalar

relativistic DFT calculations. The calculations in aqueous solvents were carried out using the

PCM solvation model. The formation of the plutonyl complexes through both monomer

agglomeration and the hydrolysis reactions of [PuO2(H2O)5]2+

were examined. The trends in the

Page 165: Relativistic Quantum Chemistry Applied to Actinides

143

calculated reaction free energies for the formation of uranyl, neptunyl and plutonyl hydroxo

complexes were discussed. The variation of the electronic structures of the complexes with

increasing number of equatorial hydroxo ligands or increasing number of plutonyl groups were

also examined.

The calculated structures of the monomeric plutonyl aquo-hydroxo complexes are in

good agreement with previously reported experimental data. The Pu-Oyl bonds become

progressively weaker and longer as the number of equatorial hydroxo ligands is increased.

Examination of the electronic structure of these complexes revealed the origin of this

phenomenon to be a decrease in the Pu-6d contributions to the plutonyl π(d) orbitals and an

increase in Pu-6d contributions to orbitals with plutonyl-σ(d)/σ(Pu-OH) and plutonyl-

σ(f)/σ/π(Pu-OH) characters. This quantitative explanation of the weakening of the Pu-Oyl bonds

supports previous speculations regarding a Pu-6d overloading or competition between the axial

Oyl and equatorial hydroxos. Similar arguments are applicable to other actinide systems. The

trend in weaker Pu-Oyl bonds results in lower plutonyl stretching vibrational frequencies. We

were able to assign the symmetric stretching modes of [PuO2(H2O)5]2+

and [PuO2(H2O)4(OH)]+

to available Raman data. The Pu-Oyl and Pu-OH distances in the plutonyl aquo-hydroxo

complexes are generally about 0.03-0.04 Å and 0.01-0.02 Å shorter than similar bonds in their

uranyl and neptunyl analogs respectively. The origin of these slight contractions is an increase in

actinyl 5f participation in the σ(f) and π(f) orbitals down the U-Np-Pu series as well as increased

stabilization of the σ(f) actinyl MO relative to the σ(d) MO.

The calculated reaction energies for the hydrolysis of [PuO2(H2O)5]2+

reveals the

formation of the dimer complex, [(PuO2)2(OH)2(H2O)6]2+

, is significantly more exothermic than

the formation of the trimer complexes. This is also the case for the uranyl and neptunyl systems.

Page 166: Relativistic Quantum Chemistry Applied to Actinides

144

The formation of the uranyl trimer was however calculated to be about 6.2 kcal/mol more

exothermic than the formation of the counterpart plutonyl complex. This ΔΔEreaction is

sufficiently accurate due to systematic cancellation of errors and in good agreement with

previously reported experimental value of 6.32 kcal/mol. The greater oxophilicity of U(VI)

means that the dehydration of two hydroxo groups in [(UO2)3(OH)5]+ to yield a μ3-oxo group in

[(UO2)3(O)(OH)3]+, is significantly more exothermic than for its plutonyl analogue. In addition,

trimeric complexes with bridging aquo ligands are more stable for the plutonyl system than for

the uranyl system. As these complexes are most likely involved in the fragmentation of the

trimeric core, this is possibly the origin of the increased mole-fraction of the trimeric and higher

polynuclear complexes in the uranyl hydrolysis speciation diagram. Comparison of the electronic

structures of [(UO2)3(O)(OH)3]+, [(NpO2)3(O)(OH)3]

+ and [(PuO2)3(O)(OH)3]

+, reveals a critical

difference in the degree of participation of the μ3-oxo 2p orbitals to the actinyl π(d)/μ3O-2p

orbitals. The μ3-oxo 2p contributions are highest in the U complex and lowest in the Pu complex.

For this reason, the π(f/d)/σ(Pu-μ3O) orbitals are also least stabilized in the Pu complex.

The calculated frequencies of the plutonyl symmetric stretching vibrational modes in the

dimer complex, [(PuO2)2(OH)2(H2O)6]2+

, match previously reported Raman spectroscopic data.

The accuracy of the calculated frequencies was ascertained by accurately predicting the Raman

and IR stretching modes of [UO2(H2O)5]2+

, [UO2(H2O)4(OH)]+, [(UO2)2(OH)2(H2O)6]

2+, and

[(UO2)3(O)(OH)3(H2O)6]+. The calculated reaction energies and the ΔΔEreaction indicate that the

mole-fraction of the uranyl trimer should exceed that of the plutonyl trimer. At high plutonyl

concentrations, the increase in the amount of trimer formed should correlate with the Raman

peak at 805 cm-1

.

Page 167: Relativistic Quantum Chemistry Applied to Actinides

145

The electronic structure of [(PuO2)2(OH)2]2+

is very similar to that of the monomer

plutonyl hydroxo species. The various π(d), π(f), σ(d) and σ(f) orbitals are linear combinations of

contributions from the two plutonyl groups. This is also the case in the trimer complex,

[(PuO2)3(OH)3(O)]+, although some MOs (especially π orbitals) contain significant overlap with

2p contributions from the central μ3-oxo atom.

References

1. Clark, D. L. H., S. S.; Jarvinen, G. D.; Neu, M. P. , The Chemistry of the Actinide and

Transactinide Elements. 3rd ed.; Springer: Dordrecht, The Netherlands, 2006; Vol. 2.10., Pg

815.

2. Runde, W.; Reilly, S. D.; Neu, M. P., Geochim. Cosmochim. Acta 1999, 63 (19-20),

3443-3449.

3. Austin, J. P.; Sundararajan, M.; Vincent, M. A.; Hillier, I. H., Dalton T. 2009, (30),

5902-5909.

4. Balasubramanian, K.; Chaudhuri, D., Chem. Phys. Lett. 2008, 450 (4-6), 196-202.

5. Chaudhuri, D.; Balasubramanian, K., Chem. Phys. Lett. 2004, 399 (1-3), 67-72.

6. Clark, D. L.; Hobart, D. E.; Neu, M. P., Chem. Rev. 1995, 95 (1), 25-48.

7. Madic, C.; Begun, G. M.; Hobart, D. E.; Hahn, R. L., Inorg. Chem. 1984, 23 (13), 1914-

1921.

8. Rao, L. F.; Tian, G. X.; Di Bernardo, P.; Zanonato, P., Chem-Eur. J. 2011, 17 (39),

10985-10993.

9. Reilly, S. D.; Neu, M. P., Inorg. Chem. 2006, 45 (4), 1839-1846.

10. Aberg, M., Acta Chem. Scand. 1970, 24 (8), 2901-&.

11. Ahrland, S., Acta Chem. Scand. 1949, 3 (4), 374-400.

Page 168: Relativistic Quantum Chemistry Applied to Actinides

146

12. Ahrland, S.; Hietanen, S.; Sillen, L. G., Acta Chem. Scand. 1954, 8 (10), 1907-1916.

13. Cao, Z. J.; Balasubramanian, K., J. Chem. Phys. 2005, 123 (11).

14. Choppin, G. R.; Mathur, J. N., Radiochim. Acta 1991, 52-3, 25-28.

15. Gianguzza, A.; Milea, D.; Millero, F. J.; Sammartano, S., Mar. Chem. 2004, 85 (3-4),

103-124.

16. Kato, Y.; Meinrath, G.; Kimura, T.; Yoshida, Z., Radiochim. Acta 1994, 64 (2), 107-111.

17. Meinrath, G.; Kato, Y.; Yoshida, Z., J. Radioanal. Nucl. Ch. 1993, 174 (2), 299-314.

18. Moll, H.; Reich, T.; Szabo, Z., Radiochim. Acta 2000, 88 (7), 411-415.

19. Neuefeind, J.; Soderholm, L.; Skanthakumar, S., J. Phys. Chem. A 2004, 108 (14), 2733-

2739.

20. Nguyen-Trung, C.; Palmer, D. A.; Begun, G. M.; Peiffert, C.; Mesmer, R. E., J. Solution

Chem. 2000, 29 (2), 101-129.

21. Palmer, D. A.; NguyenTrung, C., J. Solution Chem. 1995, 24 (12), 1281-1291.

22. Quiles, F.; Burneau, A., Vib. Spectrosc. 2000, 23 (2), 231-241.

23. Sonnenberg, J. L.; Hay, P. J.; Martin, R. L.; Bursten, B. E., Inorg. Chem. 2005, 44 (7),

2255-2262.

24. Toth, L. M.; Begun, G. M., J. Phys. Chem. 1981, 85 (5), 547-549.

25. Tsushima, S.; Yang, T. X.; Suzuki, A., Chem. Phys. Lett. 2001, 334 (4-6), 365-373.

26. Zanonato, P.; Di Bernardo, P.; Bismondo, A.; Liu, G. K.; Chen, X. Y.; Rao, L. F., J. Am.

Chem. Soc. 2004, 126 (17), 5515-5522.

27. Batista, E. R.; Martin, R. L.; Hay, P. J., J. Chem. Phys. 2004, 121 (22), 11104-11111.

28. Cao, Z. J.; Balasubramanian, K., J. Chem. Phys. 2005, 123 (11).

Page 169: Relativistic Quantum Chemistry Applied to Actinides

147

29. Clavaguera-Sarrio, C.; Vallet, V.; Maynau, D.; Marsden, C. J., J. Chem. Phys. 2004, 121

(11), 5312-5321.

30. Craw, J. S.; Vincent, M. A.; Hillier, I. H.; Wallwork, A. L., J. Phys. Chem. 1995, 99 (25),

10181-10185.

31. Horowitz, S. E.; Marston, J. B., J. Chem. Phys. 2011, 134 (6).

32. Ismail, N.; Heully, J. L.; Saue, T.; Daudey, J. P.; Marsden, C. J., Chem. Phys. Lett. 1999,

300 (3-4), 296-302.

33. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2011, 115 (48), 14110–14119.

34. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109 (48), 10961-10974.

35. Conradson, S. D.; et al, Inorg. Chem. 2004, 43 (1), 116-131.

36. ADF2012.01, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands,

http://www.scm.com.

37. te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.; Fonseca Guerra, C.; Baerends,

E. J.; Snijders, J. G.; Ziegler, T., J. Comput. Chem. 2001, 22, 931-967.

38. Frisch, M. J.; et al Gaussian 03, Revision C.02. 2004.

39. Laikov, D. N.; Ustynyuk, Y. A., Russ. Chem. B+ 2005, 54 (3), 820-826.

40. Dyall, K. G., J. Chem. Phys. 1994, 100 (3), 2118-2127.

41. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2010, 114 (4), 1957-1963.

42. Shamov, G. A.; Schreckenbach, G.; Vo, T. N., Chem-Eur. J. 2007, 13 (17), 4932-4947.

43. Bridgeman, A. J.; Cavigliasso, G.; Ireland, L. R.; Rothery, J., J. Chem. Soc. Dalton 2001,

(14), 2095-2108.

44. Faas, S.; Snijders, J. G.; van Lenthe, J. H.; van Lenthe, E.; Baerends, E. J., Chem. Phys.

Lett. 1995, 246 (6), 632-640.

Page 170: Relativistic Quantum Chemistry Applied to Actinides

148

45. van Lenthe, E., J. Comp. Chem. 1999, 20 (1), 51-62.

46. van Lenthe, E.; Baerends, E. J.; Snijders, J. G., J. Chem. Phys. 1993, 99 (6), 4597-4610.

47. Pye, C. C.; Ziegler, T., Theor. Chem. Acc. 1999, 101 (6), 396-408.

48. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74 (6), 1245-1263.

49. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100 (10), 7535-7542.

50. Miertus, S.; Scrocco, E.; Tomasi, J., Chem. Phys. 1981, 55 (1), 117-129.

51. Odoh, S. O.; Walker, S. M.; Meier, M.; Stetefeld, J.; Schreckenbach, G., Inorg. Chem.

2011, 50 (7), 3141-3152.

52. Ingram, K. I. M.; Haller, L. J. L.; Kaltsoyannis, N., Dalton T. 2006, (20), 2403-2414.

53. Tsushima, S.; Rossberg, A.; Ikeda, A.; Muller, K.; Scheinost, A. C., Inorg. Chem. 2007,

46 (25), 10819-10826.

54. Quiles, F.; Chinh, N. T.; Carteret, C.; Humbert, B., Inorg. Chem. 2011, 50 (7), 2811-

2823.

55. da Silva, G.; Kennedy, E. M.; Dlugogorski, B. Z., J. Phys. Chem. A 2006, 110 (39),

11371-11376.

56. Jang, Y. H.; Goddard, W. A.; Noyes, K. T.; Sowers, L. C.; Hwang, S.; Chung, D. S., J.

Phys. Chem. B 2003, 107 (1), 344-357.

57. Liptak, M. D.; Shields, G. C., J. Am. Chem. Soc. 2001, 123 (30), 7314-7319.

58. Di Bernardo, P.; Zanonato, P.; Bismondo, A.; Jiang, H. J.; Garnov, A. Y.; Jiang, J.; Rao,

L. F., Eur. J. Inorg. Chem. 2006, (22), 4533-4540.

59. Hummel, W.; Puigdomenech, I.; Rao, L. F.; Tochiyama, O., Comptes Rendus Chimie

2007, 10 (10-11), 948-958.

60. Rao, L. F., Chem. Soc. Rev. 2007, 36 (6), 881-892.

Page 171: Relativistic Quantum Chemistry Applied to Actinides

149

61. Reed, W. A.; Rao, L. F.; Zanonato, P.; Garnov, A. Y.; Powell, B. A.; Nash, K. L., Inorg.

Chem. 2007, 46 (7), 2870-2876.

62. Tian, G. X.; Rao, L. F., J. Chem. Thermodyn 2009, 41 (4), 569-574.

63. Clark, D. L.; Conradson, S. D.; Donohoe, R. J.; Keogh, D. W.; Morris, D. E.; Palmer, P.

D.; Rogers, R. D.; Tait, C. D., Inorg. Chem. 1999, 38 (7), 1456-1466.

Page 172: Relativistic Quantum Chemistry Applied to Actinides

150

Preface to Chapter 5

This chapter is based on a manuscript submitted to the journal “Inorganic Chemistry”. The full

citation of the paper is as follows:

Samuel O. Odoh and Georg Schreckenbach, “DFT Study of Uranyl Peroxo Complexes with

H2O, F-, OH

-, CO3

2- and NO3

-”, Inorg. Chem., 2012, submitted.

In this chapter, the structures and electronic properties of complexes formed between uranyl,

peroxo and other environmentally relevant ligands were studied using density functional theory.

At high peroxide concentrations, actinyl species form polynuclear crystalline species. These

species have been extensively studied using both experimental and theoretical techniques. The

smaller uranyl peroxo complexes formed at lower concentrations with other ligands might be

more relevant to the migration of radionuclides in the environment. In addition, a study of such

complexes might provide insights into the origin (mechanism of formation) of the larger

polynuclear uranium peroxides. Such a study is compiled in this chapter.

All the calculations in the published manuscript and compiled in this chapter were carried out by

Samuel O. Odoh. The manuscript was prepared together with Prof. Georg Schreckenbach.

Copyright permissions have been obtained from the American Chemical Society and the other

authors.

Page 173: Relativistic Quantum Chemistry Applied to Actinides

151

Chapter 5: DFT Study of Uranyl Peroxo Complexes with H2O, F-,

OH-, CO3

2- and NO3

-

Abstract

The structural and electronic properties of monoperoxo and diperoxo uranyl complexes

with aquo, fluoride, hydroxo, carbonate and nitrate ligands have been studied using scalar

relativistic density functional theory (DFT). Only the complexes in which the peroxo ligands are

coordinated to the uranyl moiety in a bidentate mode were considered. The calculated binding

energies confirm that the affinity of the peroxo ligand for the uranyl group far exceeds that of the

F-, OH

-, CO3

2-, NO3

- and H2O ligands. The formation of the monoperoxo complexes from

UO2(H2O)52+

and HO2- were found to be exothermic in solution. In contrast, the formation of the

monouranyl-diperoxo, UO2(O2)2X24-

or UO2(O2)2X4-/3-

complexes were all found to be

endothermic in aqueous solution. This suggests that the monoperoxo species are the terminal

mono-uranyl peroxo complexes in solution, in agreement with recent experimental work.

Overall, we find that the properties of the uranyl-peroxo complexes conform to well known

trends: the coordination of the peroxo ligand weakens the U-Oyl bonds, stabilizes the σ(d)

orbitals and causes a mixing between the uranyl π- and peroxo σ and π orbitals. The weakening

of the U-Oyl bonds upon peroxide coordination results in uranyl stretching vibrational

frequencies that are much lower than those obtained after the coordination of carbonato or

hydroxo ligands.

Page 174: Relativistic Quantum Chemistry Applied to Actinides

152

Introduction

The coordination of actinide ions to the peroxide group, O22-

, has been highlighted for use

in nuclear separation technologies due to the crystallization of actinide peroxides complexes1, in

addition to the intensification of the corrosion of uranium dioxide nuclear fuels after peroxide-

induced oxidation.2-3

The peroxide group has a very strong affinity for uranium resulting in

insoluble polynuclear solids at high concentrations and pH.1, 4-7

In fact, studtite, UO4.4H2O and

metastudtite, UO4.2H2O, two hydrated uranyl peroxides, are the only known peroxide containing

minerals and are formed from hydrogen peroxide generated by the α-radiolysis of water.8

Recently, the agglomeration of uranyl-peroxide units into nanoscale cage clusters has been the

focus of several studies. Burns et al. have synthesized and characterized a variety of uranyl

peroxide hydroxide polyhedral species.1, 4-7, 9

These polynuclear clusters mostly adopt a cage-like

motif with one of the largest of them being a [UO2(OH)(O2)]6060-

complex which adopts the

buckyball structure of Buckminsterfullerene. The reason behind the preference of the cage motif

over the linear sheet motif in these and other uranyl-peroxide nanoclusters has been investigated

by several workers using electronic structure calculations.10-11

In their calculations Vlaisavljevich

et al. showed that the uranyl-peroxide-uranyl motif found in these clusters is inherently bent as a

result of the covalent interaction across the U-O2 bond.11

They also demonstrated the effect of

the size and electronegativity of the counterion on the dihedral U-O2-U angle.

It is however the case that the solution chemistry of uranyl peroxide has not been

investigated to the same extent as its solid-state complexes. The solution chemistry of these

peroxide complexes is particularly important in view of the environmental importance of

migrating nuclear waste streams. Goff et al.12

have shown that minute amounts of peroxide can

be used to displace a carbonate group from aqueous solutions of UO2(CO3)34-

. Structural and

Page 175: Relativistic Quantum Chemistry Applied to Actinides

153

spectroscopic characterization of the peroxo-carbonato complex formed revealed it to be

UO2(O2)(CO3)24-

in the form of K4[UO2(O2)(CO3)2].H2O. The spectroscopic investigations of

Meca et al. have also indicated the existence of two uranyl peroxide-hydroxide complexes at pH

12 in the absence of carbonate species.13

These species were suggested to be UO2(O2)(OH)22-

and UO2(O2)2(OH)24-

, with the latter, as expected, being more predominant at higher peroxide

concentrations. More recently, Zanonato et al. examined the ternary peroxide-hydroxide system

in a tetramethylammonium nitrate medium.14

They found UO2(O2)(OH)- to be the dominant

complex from pH 9.5 to 11.5 even though significant amounts of the binuclear complex,

(UO2)2(O2)2(OH)-, was present at around pH 10.5. These studies suggest that the uranyl peroxide

moiety, UO2(O2) could indeed form complexes with a wide variety of ligands in so far as the

U(VI) and O22-

concentration is controlled, to prevent the complexity and precipitation

engendered by polynuclear species. Nyman et al. have also recently reported the synthesis of two

lithium salts of UO2(O2)34-

.15

They suggested that these monomeric triperoxo systems could play

a role in aqueous behavior, re-dissolution and self assembly characteristics of uranyl polynuclear

peroxides.

The recent progress being made in the solvent phase and monomer chemistry of uranyl-

peroxide species has motivated us to carry out a systematic computational study of the structural

and electronic properties of the possible ternary uranyl peroxide complexes with fluoride, aquo,

hydroxo, carbonate and nitrate ligands. Additionally, we have focused on the characteristic

vibrational frequencies of these peroxo complexes, the relative stabilities of their various

structures as well as the trends in their calculated structural properties. Ultimately the aim of the

current work is to provide calculated structural data for the uranyl peroxo complexes. This will

hopefully allow for an easier characterization of some these complexes should they be

Page 176: Relativistic Quantum Chemistry Applied to Actinides

154

synthesized in the future. All the calculations in this work have been carried out using scalar

relativistic density functional theory (DFT). The use of DFT calculations as a complement to the

available experimental stoichiometric and structural data of actinide complexes and in predicting

the structures of as yet to be synthesized actinide species is well established.11, 16-41

Computational Details

The geometry optimizations in this work were all carried out with the Gaussian 03

code.42

Vibrational frequency analyses with the harmonic approximation were carried out after

the geometry optimizations to characterize the local minima nature of the optimized structures on

the potential energy surfaces. Most of the geometry optimizations and vibrational frequency

calculations were carried out in aqueous solution while employing the polarizable continuum

solvation (PCM) model.43

In some cases, the gas phase structures of the complexes were also

optimized and presented in this work, for comparison. In the PCM calculations, the default

atomic radii of the united force field (UFF) in Gaussian 03 were employed. The peroxide

complexes studied in this work were all found to be singlet species. The restricted singlet

wavefunctions are stable with respect to conversion to unrestricted determinants and relaxation

of the orbital symmetries. In all the calculations, only complexes with bidentate coordination

between the uranium atoms and the peroxo group were considered. Ultra-fine grids were used in

the numerical integration of the exchange-correlation portion of the density functional.

Regarding basis sets, the Stuttgart small-core scalar-relativistic pseudopotential was used to

describe the uranium atoms.44-46

The pseudopotential represents 60 core electrons in uranium

while the remaining 32 electrons were represented by the associated valence basis set. The

design and use of this pseudopotential-basis set combination reduces the computational expense

and allows a wise inclusion of scalar-relativistic effects. To further improve computational

Page 177: Relativistic Quantum Chemistry Applied to Actinides

155

efficiency, all g-type functions were removed from the valence basis set. The 6-31+G* basis was

used to describe oxygen, nitrogen, fluorine and carbon atoms while hydrogen atoms were

described with the 6-31G basis. We label this scheme in which the uranium RECP and the 6-

31+G*/6-31G bases for the non-actinide atoms were employed as B1. The B3LYP47-49

functional

was employed in all the calculations carried out in this current work. Overall the B3LYP/B1

abbreviation is used to describe the functional and basis set combination employed in the

calculations.

Single point calculations on the geometries optimized at the B3LYP/B1 level were

carried out with a four-component scalar relativistic approach as implemented in the Priroda

program.50

These calculations allowed us to obtain the population based Mayer bond orders

which are in our experience good reflections of the formal bond order.34, 37

A basis of double-ζ

quality (cc-pVDZ) was used for all the elements for the large component with the corresponding

kinetically-balanced basis sets for the small component.51

The B3LYP functional was also

employed in the Priroda calculations. This combination of functional and basis set is labeled as

B3LYP/B2.

Results and Discussions

UO22+

and its peroxo derivatives. The electronic structure of UO22+

as well as those of its

neptunium and plutonium analogues has been studied extensively.52-57

At the B3LYP/B1 level

and in aqueous solution, the valence region of this dication consists of the σ(f), π(f), σ(d) and

π(d) orbitals at -11.1, -11.9, -12.2 and -12.2 eV respectively. The optimized geometries for

UO22+

and its peroxide derivatives, UO2(O2)n2-n

, obtained with the PCM solvation model, are

shown in Figure 5.1. For the uranyl peroxide species, UO2(O2), UO2(O2)22-

and UO2(O2)34-

, the σ

Page 178: Relativistic Quantum Chemistry Applied to Actinides

156

Figure 5.1: The structures of UO22+

and its peroxo derivatives optimized at the B3LYP/B1 level

in aqueous solution. The D2h structure of the uranyl diperoxide was found to be a transition state

structure.

and π bonding orbitals of the peroxo ligand are found below the uranyl σ(d) and π(d) orbitals,

Table 5.1. There is however substantial mixing between the O2-π orbital and the uranyl π(d)

orbitals in UO2(O2). In contrast, the U-Operoxo bond orbitals between the uranyl and peroxo units

are found above the uranyl σ(f) orbitals. The uranyl and peroxo orbitals of UO2(O2) are shown in

Figure 5.2. The coordination of the second and third peroxide ligands stabilizes the actinyl π(f)

orbitals such that there is significant σ(O2)-π(f) as well as π(O2)-π(d) mixing in the diperoxo and

triperoxo complexes, Table 5.1. In addition, the uranyl π(f) orbitals are stabilized below the π(d)

Page 179: Relativistic Quantum Chemistry Applied to Actinides

157

orbitals in UO2(O2)34-

. The nature of the π(O2)-π(d) mixing is reminiscent of the π(d)-μ3O2(2p)

orbitals that were recently reported in trimeric [(AnO2)3(O)(OH)3]+ complexes of uranium and

plutonium.58

In that study, the degree of stability conferred by the central μ3-oxo ligand on the

hexagonal trimer shape was found to be greater for the uranium complex and lower for the

plutonium complex. A similar difference in the mixing of the σ(O2) and π(O2) orbitals with the

uranyl and plutonyl π(f) and π(d) orbitals might also exist for the peroxide complexes. The C2v

structure of UO2(O2)22-

was calculated to be 3.1 kcal/mol more stable than its D2h conformer in

the aqueous phase. The energy difference in the gas phase was calculated as 1.2 kcal/mol. The

D2h structure is a transition state structure in both gaseous and aqueous media as it possesses

imaginary frequencies of about 37i cm-1

and 76i cm-1

respectively in these media. These

imaginary frequencies correspond to the equatorial bending of the O2-U-O2 group. The bent O2-

U-O2 wing in the C2v structure suggests a preference for circular polynuclear (UO2(O2)2)n-type

species over the linear polymer species implied by the D2h structure.

The calculated geometries of UO22+

, UO2(O2), UO2(O2)22-

and UO2(O2)34-

obtained at the

B3LYP/B1 level in aqueous solution are presented in Table 5.2. The calculated symmetric and

asymmetric vibrational stretching modes of the uranyl groups in these complexes as well as the

O2 stretching and U-O2 stretching modes of the peroxides are also presented in Table 5.2. The

Oyl-U-Oyl bond angle was calculated as 180.0°, 174.3°, 171.6° and 180.0° in UO22+

, UO2(O2),

C2v-UO2(O2)22-

and UO2(O2)34-

respectively. There is a sequential weakening of the U-Oyl, U-

Operoxo, O-Operoxo bonds as the number of coordinated peroxide ligands are increased. The weaker

U-Oyl bonds result in lower uranyl stretching vibrational frequencies. The uranyl stretching

modes in UO2(O2)34-

were calculated as 662 cm-1

for the asymmetric mode and 646 cm-1

for the

symmetric mode. These are significantly lower than those obtained for the bare dication. To put

Page 180: Relativistic Quantum Chemistry Applied to Actinides

158

Table 5.1: Energies and characters of the MOs of the dioxouranium (VI) peroxides in aqueous

solution obtained at the B3LYP/B1 level.

UO2(O2) UO2(O2)22-

UO2(O2)34-

MO Energy

(eV)

Character MO Energy

(eV)

Character MO Energy

(eV)

Character

22 -11.32 σ(O2) 26 -10.07 σ(O2) 30 -9.19 σ(O2)

23 -11.05 π(O2)/π(d) 27 -9.77 σ(O2) 31 -8.89 σ(O2)/σ(d)

24 -10.77 π(O2)/π(d) 28 -9.64 π(O2)/σ(d) 32 -8.83 σ(O2)

25 -9.84 σ(d) 29 -9.58 π(O2)/π(d) 33 -8.82 σ(O2)

26 -9.84 π(d) 30 -9.38 π(O2)/π(d/f) 34 -8.50 π(O2)/π(d)

27 -9.35 π(f) 31 -8.75 π(O2)/π(f) 35 -8.50 π(O2)/π(d)

28 -9.35 π(f) 32 -8.35 σ(d) 36 -8.13 π(O2)/σ(f)

30 -8.11 σ(f) 33 -8.07 π(d/f) 37 -7.75 π(O2)/π(d/f)

31 -7.40 σ(U-O2) 34 -7.69 π(f) 38 -7.75 π(O2)/π(d/f)

35 -7.51 π(d) 39 -7.28 σ(d)

36 -7.42 π(f) 40 -6.50 π(f)

37 -6.57 σ(f) 41 -6.50 π(f)

38 -6.03 σ(U-O2) 42 -6.42 π(d)

43 -6.42 π(d)

44 -5.61 σ(f)

45 -5.06 σ(U-O2)

Page 181: Relativistic Quantum Chemistry Applied to Actinides

159

Figure 5.2: The molecular orbitals of UO2(O2). The geometry of this complex was optimized

with the PCM approach and the B3LYP functional.

this in perspective, the uranyl stretching modes in UO2(NO3)3- and UO2(CO3)3

4-, two complexes

also possessing bidentate anionic ligands, were calculated as 870 and 929 cm-1

for the former and

775 and 811 cm-1

in the latter. In addition, the longer U-Operoxo and O-Operoxo bonds result in

lower frequencies for the O2 stretching and U-O2 stretching vibrational modes respectively.

Similar elongations of the U-Oyl bonds have been observed in the uranyl hydroxide, UO2(OH)n2-

n,58-59

and fluoride, UO2Fn2-n

,29

complexes. The simulated IR spectra of UO22+

, UO2(O2), C2v-

UO2(O2)22-

and UO2(O2)34-

are shown in Figure 5.3. The vibrational modes associated with the

stretching of the peroxo O-O bonds were found between 839 and 936 cm-1

, Table 5.2, gradually

Page 182: Relativistic Quantum Chemistry Applied to Actinides

160

decreasing down the series and split across several frequencies in the diperoxo and triperoxo

species. The calculated IR intensities of these peaks indicate that some of the O-O stretching

should be observed in the Raman spectra of the peroxo complexes. The peak centered at around

470 cm-1

in the simulated IR spectra of UO2(O2) and around 400 cm-1

in C2v-UO2(O2)22-

and

UO2(O2)34-

, Figure 5.3, contains the U-Operoxo stretching vibrations of which the symmetric mode

has significantly higher IR intensities than the counterpart asymmetric mode. Similar to the case

for the O-O stretching mode, the U-Operoxo stretching vibrations are split into IR and Raman

active modes in the higher peroxides. The U-Oyl bonds in UO2(O2)34-

optimized with the PCM

Figure 5.3: Simulated IR spectra of UO22+

and its peroxo derivatives obtained at the at the

B3LYP/B1 level in aqueous solution.

Page 183: Relativistic Quantum Chemistry Applied to Actinides

161

Table 5.2: Calculated structural properties and vibrational frequencies of UO22+

and its peroxo

derivatives obtained at the B3LYP/B1 level in aqueous solution.

Bonds (Å) Vibrational Frequencies (cm-1

)

U-Oyl U-O

peroxo

O-O

peroxo

Uranyl Stretching O2 stretch U-O2 stretch

Asymm. Sym. Symm. Asymm.

UO22+

1.748 1002 920

UO2(O2) 1.810 2.177 1.442 854 789 936 472 464

UO2(O2)22-

C2v 1.866 2.233/

2.260

1.471 746 712 891/851 434/394 406/394

D2h 1.861 2.259 1.467 755 713 891/856 393/392 397/370

UO2(O2)34-

1.907 2.323 1.485 662 646 871/841/

839

378/354/

332

377/339/3

28

Expt. 1.846a 2.303-

2.324a

a X-ray structure of Li4[UO2(O2)3]3.10H2O; Reference

15

model are about 0.05 Å longer than those found in solid Li4[UO2(O2)3]3.10H2O, Table 5.2.15

It is

noted that the crystal structure indicated interactions between the lithium-bound water molecules

and the peroxide oxygen ligands. In implicit solvation models, like the PCM model employed

here, the effect of a solvent is included with a statistically average solvent described by its

Page 184: Relativistic Quantum Chemistry Applied to Actinides

162

dielectric constant, a macroscopic property. These models are not sufficient to describe the

lithium-water and water-peroxo interactions in Li4[UO2(O2)3]3.10H2O.

There are two types of U-Operoxo bonds in the C2v structure of UO2(O2)22-

, the proximal

ones being about 0.03 Å longer than the distal ones, Figure 5.1. The σ(O2) character of MO-26 in

C2v-UO2(O2)22-

is such that there is some overlap across the distal oxygen atoms of the two

peroxo ligands, Figure 5.4. As this overlap is prohibited in the D2h structure, by virtue of the

trans arrangement of the equatorial peroxo groups, it is most likely the reason behind the greater

stability of the C2v structure. Examination of the MO energy levels reveals that MO 26 is about

0.30 eV (6.92 kcal/mol) below MO 27 (σ orbital antibonding across the O-O′ distance, Figure

5.3) in C2v-UO2(O2)22-

. In contrast, the energy difference between these orbitals is 0.08 eV (1.84

kcal/mol) in D2h-UO2(O2)22-

. This discrepancy indicates a stabilization of MO-26 in the C2v

structure. The difference in the relative energies of these orbitals is however not sufficient to

fully explain the greater stability of C2v-UO2(O2)22-

. Increased contributions from uranium

atomic orbitals, generally found in D2h-UO2(O2)22-

, result in destabilization of molecular orbitals

with predominantly peroxo character. The converse is true for orbitals with mainly uranyl

character.

The σ(U-O2) orbitals of the U-Operoxo bonds are formed by overlap of the uranium 5f

orbitals and the in-plane π antibonding orbitals of the peroxo ligand, Figure 5.2. There is some

U-6d contribution, 8%, to the σ(U-O2) orbital in UO2(O2). The U-6d orbitals however do not

participate in the U-Operoxo bonds of UO2(O2)22-

and UO2(O2)34-

. The out-of-plane peroxo

antibonding π orbitals are found at higher energies.

Page 185: Relativistic Quantum Chemistry Applied to Actinides

163

The absolute ligand binding energies for UO2(O2), C2v-UO2(O2)22-

and UO2(O2)34-

were

calculated as -121.5, -194.4 and -235.8 kcal/mol respectively at the B3LYP/B1 level in aqueous

solution. These are the energies required to bind the peroxo ligands to the UO22+

and are

equivalent to -121.5, -97.2 and -78.6 kcal/mol per peroxo group in these complexes respectively.

The binding energies were obtained from the calculated electronic energies of the species

involved in the reaction: UO22+

(aq) + nO22-

(aq) → UO2(O2)n2-2n

(aq). Although, the absolute

binding energies become larger for successive coordination of a peroxo ligand, the calculated

binding energies relative to the (n-1) species decrease down the series. A similar effect has been

Figure 5.4: MO-26 and MO-27 of the C2v structure of UO2(O2)22-

. The former is bonding with

respect to the O-O′ distance between the distal oxygen atoms of the peroxo group while the latter

is antibonding. The distal U-O bonds are about 0.03 Å shorter than the proximal ones.

observed in our previous works on plutonyl hydroxides58

and uranyl fluorides29

as well as by

other workers. For comparison, we also calculated the absolute ligand binding energies of

UO2(OH)42-

, UO2(NO3)3- and UO2(CO3)3

4-. These were calculated as -167.8, -55.4 and -148.0

kcal/mol respectively, or -41.9, -18.5, -49.3 kcal/mol per ligand respectively. The implication of

Page 186: Relativistic Quantum Chemistry Applied to Actinides

164

this is that for the UO2Ln species (L = O22-

, OH, NO3- or CO3

2-), the affinity of the uranyl group

for the peroxide ligand far exceeds its affinity for the hydroxide, nitrate and carbonate ligands.

Uranyl aquo complexes. The uranyl chemistry in highly acidic solutions is dominated by the

UO2(H2O)52+

complex. The calculated bond lengths and vibrational frequencies of UO2(H2O)52+

obtained in this work are presented in Table 5.3. Shamov et al. and several other workers have

previously predicted the structure of this complex in the gas phase at the DFT level.16, 35-36, 59

The

calculated gas phase geometry of this complex obtained in this work is in good agreement with

previous reports. Introduction of solvent effects with the continuum solvation model slightly

weakens the U-Oyl bonds by about 0.01 Å but strengthens the U-OH2 interactions by about 0.05

Å. This is in accordance with previous computational work.28-29, 34-36

The use of the solvation

model brings the calculated U-Oyl and U-OH2 bond-lengths to within 0.01 and 0.03 Å of the

available aqueous phase extended X-ray absorption fine structure, EXAFS, data of Allen et al.60

The calculated uranyl stretching vibrational frequencies are reduced by about 20-53 cm-1

in the

solvent phase calculations. The experimental vibrational frequencies61-62

of UO2(H2O)52+

are

within 41 cm-1

of the calculated values obtained using the PCM model.

The replacement of aquo ligands in UO2(H2O)52+

by an equatorial peroxo group forms

UO2(H2O)4(O2), UO2(H2O)3(O2) or UO2(H2O)2(O2). This is particularly interesting given that the

chemical formulas of the minerals studtite and metastudtite are UO2(H2O)4(O2) and

UO2(H2O)2(O2) respectively. The optimized structures for the aquo-peroxo complexes are

presented in Figure 5.5. Starting from gas phase UO2(H2O)52+

and HO2-, as shown in the reaction

below, the formation of the aquo-peroxo complexes were calculated to be very exothermic in the

gas phase, between -158.2 kcal/mol for UO2(H2O)2(O2) and -187.1 kcal/mol for UO2(H2O)4(O2).

Page 187: Relativistic Quantum Chemistry Applied to Actinides

165

Table 5.3: Calculated structural properties and vibrational frequencies (cm-1

) of UO2(H2O)52+

and its peroxo derivatives obtained at the B3LYP/B1 level in the gas phase and in aqueous

solution.

Bond Lengths (Å) Uranyl Stretching

U-Oyl U-Oaquo U-Operoxo O-Operoxo Asymm. Symm.

UO2(H2O)52+

Gas phase 1.750 2.491 1026 937

Solution 1.759 2.440-2.492 973 911

Experimental 1.76a 2.41

a 965

b 870

b

UO2(O2)(H2O)4

Gas phase 1.813 2.677/2.686 2.176 1.426 886 813

Solution 1.815 2.644-2.666 2.195 1.440 840 789

UO2(O2)(H2O)3

Gas phase 1.814 2.593/2.600 2.162 1.429

Solution 1.814 2.549-2.557 2.186 1.445 843 791

UO2(O2)(H2O)2

Gas phase 1.808 2.571 2.141 1.435 893 820

Solution 1.813 2.562/2.612 2.179 1.444 844 788

Solution

UO2(O2)2(H2O)2-

1.866 2.727 2.240/2.264 1.471 739 704

cis-UO2(O2)2(H2O)22-

1.862 2.725 2.251/2.275 1.469 749 718

trans-UO2(O2)2(H2O)22-

1.865 2.645 2.267/2.279 1.466 741 709

a Reference

60 b References 61-62

Page 188: Relativistic Quantum Chemistry Applied to Actinides

166

Figure 5.5: The structures of UO2(H2O)52+

and its peroxo derivatives optimized at the

B3LYP/B1 level in aqueous solution.

UO2(H2O)52+

+ HO2- → UO2(O2)(H2O)n + (4-n) H2O + H3O

+

There is a drastic solvent effect, in excess of 150 kcal/mol, on these reactions. The formation of

UO2(H2O)4(O2) is still the most exothermic in the solution phase, about -13.7 kcal/mol,

compared to -6.7 and -8.7 kcal/mol for UO2(H2O)2(O2) and UO2(H2O)3(O2) respectively. This

implies that the preferred equatorial coordination number of uranyl aquo-peroxo complex in the

gas phase and in solution is 6, larger than that in the case of UO2(H2O)52+

.35-36, 63

The U-Oyl bond

lengths in the aquo-peroxo complexes are centered at 1.814 Å with minimal changes both in

solution and as the number of aquo ligands are increased from 2 to 4, Table 5.3. This represents a

weakening of about 0.055 Å from UO2(H2O)52+

. The identical U-Oyl bond lengths in these aquo-

peroxo complexes are reflected in the similar calculated uranyl stretching vibrational

frequencies, Table 5.3. The presence of the equatorial aquo ligands also has little influence on

the length of the peroxo O-O bond. As a result the O-O stretching mode of UO2(O2) calculated at

Page 189: Relativistic Quantum Chemistry Applied to Actinides

167

936 cm-1

in solution is retained in the aquo-peroxo complexes with very little change in its

frequency. The U-Operoxo bonds become slightly longer from UO2(O2)(H2O)2 to UO2(O2)(H2O)4,

weakening by 0.035 Å for the gaseous species and 0.016 Å in solution. The stretching modes

associated with the U-Operoxo bonds are centered around 460-470 cm-1

just as in UO2(O2), Figure

5.3. In contrast, the weakening of the U-OH2 bonds after coordination of an equatorial peroxo

ligand is significantly more pronounced, 0.08-0.20 Å in the gas phase and about 0.10 Å with the

PCM model. The steric crowding between the aquo and peroxo ligands results in longer U-OH2

bonds as the equatorial coordination number is increased.

It is reasonable to expect the formation of diperoxo species such as UO2(O2)2(H2O)2-

or

UO2(O2)2(H2O)22-

at high peroxide concentrations. The calculated thermochemistry for the

formation of these complexes in solution, via the reaction shown below, is however unfavorable.

The reactions are endothermic for both UO2(O2)2(H2O)2-

, 22.7 kcal/mol and UO2(O2)2(H2O)22-

,

18.1 kcal/mol. In a similar manner, the formation of the triperoxo complex, UO2(O2)34-

, from

UO2(H2O)52+

, is endothermic in aqueous solution by 77.8 kcal/mol.

UO2(H2O)52+

+ 2HO2- → UO2(O2)2(H2O)n

2- + 3-n H2O + 2H3O

+

UO2(H2O)52+

+ 3HO2- → UO2(O2)3

4- + 2H2O + 3H3O

+

Uranyl fluoride complexes. The structural properties of UO2F42-

as well as those of species

formed by substitution of the fluoride ligands by peroxo groups are presented in Table 5.4. These

geometrical features were obtained at the B3LYP/B1 level while employing the PCM solvation

model. The optimized geometries for these complexes are presented in Figure 5.6. The calculated

structural parameters of UO2F42-

are in good agreement with experimental data64

and previous

Page 190: Relativistic Quantum Chemistry Applied to Actinides

168

Table 5.4: Calculated structural properties and vibrational frequencies of UO2F42-

and its peroxo

derivatives obtained at the B3LYP/B1 level in aqueous solution.

Bonds (Å) Vibrational Frequencies (cm-1

)

U-Oyl U-O

peroxo

O-O

peroxo

U-F Stretching Modes

Uranyl O2 U-O2 stretch

Asymm. Symm. Symm. Asymm.

UO2F42-

1.826 2.216 822 786

Expt. 1.800a 2.260

a

UO2F3(O2)3-

1.859 2.258 1.461 2.302 752 720 895 405 390

UO2F2(O2)2-

1.842 2.235 1.460 2.251

2.260

787 739 891 411 400

UO2F(O2)23-

1.876 2.270-

2.284

1.477 2.328 722 692 878/

843

398 394

UO2F2(O2)24-

Cis 1.893 2.299-

2.317

1.466 2.434 682 660 904/

875

364/

340

380/

352

Trans 1.893 2.309 1.465 2.423 687 664 904

/

880

352/

345

360/

343

a Reference

40

theoretical literature29, 64-65

. The U-Oyl bonds were calculated to be about 0.03 Å longer than the

experimental bond length of 1.800 Å. In contrast, the U-F bonds are about 0.04 Å shorter than

Page 191: Relativistic Quantum Chemistry Applied to Actinides

169

the experimental value of 2.260 Å. The vibrational mode associated with the asymmetric

stretching of the uranyl group was calculated as 822 cm-1

in good agreement with the value

obtained from experiment. The experimental data on the structural properties of this complex

were obtained by EXAFS, measurements in aqueous solution.64

Structurally, the U-Oyl bond lengths increase sequentially by about 0.02 Å down the

UO2F42-

, UO2F2(O2)2-

, UO2F3(O2)3-

, UO2F(O2)23-

and UO2F2(O2)24-

series. The U-F and U-Operoxo

bonds also become progressively weaker down the series, Table 5.4. It is not entirely surprising

that the U-Oyl, U-F and U-Operoxo bonds are longer in UO2F3(O2)3-

than in UO2F2(O2)2-

, or in

UO2F2(O2)24-

than in UO2F(O2)23-

. As previously noted, this trend towards weaker bonds as the

number of equatorially coordinated anionic ligands is increased has been reported by various

workers in uranyl hydroxides59, 66

, fluorides29

, carbonates20

and nitrates20

. Clark et al. suggested

that the similar trend for the uranyl hydroxides was most likely due to competition between the

axial oxo groups and the equatorial hydroxide ligands for the uranium 6d orbitals.66

In our recent

study of the plutonyl hydroxide complexes, [PuO2(H2O)4-n(OH)n]2-n

, we found that the Pu-OH

bond is supported by both π-type and σ-type bond orbitals, with increasing Pu-6d contributions

to these orbitals as the number of equatorial hydroxide ligands is increased and simultaneous

decrease in the Pu-6d contributions to the axial plutonyl π-orbitals.58

The binding energies of the fluoro groups to the uranyl peroxo cores, UO2(O2) or

UO2(O2)22-

, were calculated to be 23.9, 20.7, 13.6 and 5.2 kcal/mol per fluoro ligand in

UO2F2(O2)2-

, UO2F3(O2)3-

, UO2F(O2)23-

and UO2F2(O2)24-

respectively. This correlates well with

the increasing U-F bond lengths down the series, Table 5.4. The weakening of the U-Oyl and U-

Operoxo bonds as the number of equatorial ligands in the ternary fluoro-peroxo complexes are

increased results in lower wavenumbers for the Oyl-U-Oyl stretching modes as well as the U-

Page 192: Relativistic Quantum Chemistry Applied to Actinides

170

Operoxo stretching modes, Table 5.4. The symmetric and asymmetric uranyl stretching modes

decrease from respectively 786 and 822 cm-1

in UO2F42-

to respectively 739 and 787 cm-1

in

Figure 5.6: The structures of UO2F42-

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution.

UO2F2(O2)2-

. The coordination of an extra fluoride ligand in the equatorial region of UO2F3(O2)3-

further reduces the frequencies of these vibrational modes to 720 and 752 cm-1

respectively,

Table 5.4. The symmetric and asymmetric U-Operoxo stretching modes also decrease from 400

and 411 cm-1

in UO2F2(O2)2-

to 390 and 405 cm-1

in UO2F3(O2)3-

. A similar pattern is also

present in the fluoro-diperoxo and difluoro-diperoxo complexes, Table 5.4. We note that

UO2F2(O2)2-

is topologically similar to the UO2(O2)(CO3)24-

ion. The asymmetric U-Operoxo

stretching mode of UO2F2(O2)2-

calculated at 400 cm-1

has a significantly lower IR intensity than

Page 193: Relativistic Quantum Chemistry Applied to Actinides

171

the symmetric mode calculated at 411 cm-1

. The calculated peak at 400 cm-1

is in a way similar

to the peak observed at 413.1 cm-1

in the Raman spectrum of K4[UO2(O2)(CO3)2].H2O.12

The reactions leading to the formation of the ternary fluoro-peroxo complexes from

UO2(H2O)52+

were studied at the B3LYP/B1 level in solution. These reactions take the form:

UO2(H2O)52+

+ HO2- + nF

- → UO2(O2)(F)n

-n + 4H2O + H3O

+ (where n = 2/3)

UO2(H2O)52+

+ 2HO2- + nF

- → UO2(O2)2(F)n

-(n+2) + 3H2O + 2H3O

+ (where n = 1/2)

The calculated energies for these reactions are -44.9, -59.3, 10.9 and 14.1 kcal/mol for

UO2F2(O2)2-

, UO2F3(O2)3-

, UO2F(O2)23-

and UO2F2(O2)24-

respectively. Firstly, the formation of

the fluoro-diperoxo species in solution is endothermic, similar to the case with the aquo-diperoxo

complexes. On the other hand, the reactions leading to the formation of the monoperoxo

complexes are significantly exothermic in solution.

Uranyl hydroxide complexes. The geometries of the hydroxo and hydroxo-peroxo analogues of

the fluoro and fluoro-peroxo complexes discussed above were also optimized at the B3LYP/B1

level in aqueous solution. The calculated structural parameters and vibrational frequencies are

presented in Table 5.5 and the structures are shown in Figure 5.7. Aqueous phase EXAFS

measurement by Moll et al. provided U-Oyl and U-OH bonds lengths of 1.83 Å and 2.26 Å

respectively for UO2(OH)42-

.67

The bond lengths obtained for these bonds from our theoretical

calculations are within 0.02 Å of the EXAFS data, Table 5.5. The frequencies of the symmetric

and asymmetric uranyl stretching vibrational modes in this complex were calculated to be 757

and 790 cm-1

respectively in solution. In the gas phase, the calculated frequencies of these

vibrational modes are 756 and 826 cm-1

respectively. The frequencies of these vibrations were

found to be 784 and 857 cm-1

in aqueous phase measurement of the IR and Raman spectra of

Page 194: Relativistic Quantum Chemistry Applied to Actinides

172

Table 5.5: Calculated structural properties and vibrational frequencies of UO2(OH)42-

and its

peroxo derivatives obtained at the B3LYP/B1 level in aqueous solution.

Bonds (Å) Vibrational Frequencies (cm-1

)

U-Oyl U-O

peroxo

O-O

peroxo

U-OH Stretching Modes

Uranyl O2 U-O2

Asymm. Symm. Symm. Asymm.

UO2(OH)42-

1.845 2.279 790 757

Expt. 1.830a 2.265

a 857

b 784

b

UO2(OH)3(O2)3-

1.871 2.281 1.464 2.342-

2.396

734 707 891 390 360

UO2(OH)2(O2)2-

1.851 2.247 1.468 2.286 774 731 872 419 401

UO2(OH)(O2)23-

1.879 2.275

2.287

1.481 2.344 718 689 870

836

393

384

378

357

UO2(OH)2(O2)24-

Cis 1.897 2.314

2.330

1.471 2.472 682 664 894

869

370

349

341

Trans 1.898 2.318 1.469 2.479 680 663 891

872

359

336

355

330

a Reference

67

b Reference

68-69

Page 195: Relativistic Quantum Chemistry Applied to Actinides

173

Figure 5.7: The structures of UO2(OH)42-

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution.

uranyl hydroxides by Quiles et al.68-69

Clark et al. assigned the Raman peak at 796 cm-1

to the

symmetric uranyl stretching mode in their characterization of a cobalt salt of the tetrahydroxide,

[Co(NH3)6]2[UO2(OH)4]3.H2O.66

Overall, it appears that the calculated vibrational frequencies

obtained in the aqueous phase deviate from the experimental values by 30-70 cm-1

. Compared to

the UO2F42-

complex, Table 5.4, the U-Oyl and U-ligand (U-OH/U-F) bonds are about 0.02 and

0.06 Å longer in the tetrahydroxo complex, Table 5.5. The origin of this bond weakening in the

hydroxo complex is most likely the differences in the extents to which the 2p atomic orbitals of

the equatorial ligands can compete with the oxo 2p orbitals for the U-6d orbitals. This is related

to the π-donating abilities of the OH- and F

- ligands. The longer U-Oyl bonds in UO2(OH)4

2- is

correlated with lower frequencies for the uranyl stretching modes in comparison to UO2F42-

.

Page 196: Relativistic Quantum Chemistry Applied to Actinides

174

The reactions leading to the formation of UO2(OH)2(O2)2-

, UO2(OH)3(O2)3-

,

UO2(OH)(O2)23-

and UO2(OH)2(O2)24-

are similar to those written for the analogous fluoro-

peroxo complexes:

UO2(H2O)52+

+ HO2- + nOH

- → UO2(O2)(OH)n

-n + 4H2O + H3O

+ (where n = 2/3)

UO2(H2O)52+

+ 2HO2- + nOH

- → UO2(O2)2(OH)n

-(n+2) + 3H2O + 2H3O

+ (where n = 1/2)

The reaction energies for these are -57.7, -69.3, 6.3 and 8.1 kcal/mol for UO2(OH)2(O2)2-

,

UO2(OH)3(O2)3-

, UO2(OH)(O2)23-

and UO2(OH)2(O2)24-

respectively in aqueous solution. The

formation of the ternary hydroxo-peroxo complexes is significantly less exothermic than the

formation of the analogous fluoro-peroxo complexes. However like the aquo-diperoxo and

fluoro-diperoxo complexes, the formation of the hydroxo-diperoxo species from HO2- was

calculated to be endothermic in solution.

The optimized structures of the hydroxo-peroxo complexes are presented in Figure 5.7.

The calculated U-Oyl bond lengths in these complexes, Table 5.5, are generally within 0.01 Å of

those obtained for their fluoride counterparts, Table 5.4. As a result of this similarity in U-Oyl

bond lengths, the calculated uranyl stretching vibrational frequencies of the hydroxo-peroxo

species are generally within 20 cm-1

of those of the fluoro-peroxides. The U-Operoxo bonds were

calculated to be about 2.247 and 2.281 Å long in UO2(OH)2(O2)2-

and UO2(OH)3(O2)3-

respectively, Table 5.5. These can be compared to 2.235 and 2.258 Å for the difluoro and

trifluoro monoperoxides, Table 5.4. The calculated U-Operoxo stretching modes are found in the

range 330-419 cm-1

, with the diperoxo species possessing two symmetric and two asymmetric U-

Operoxo stretching modes. This is similar to the case in the analogous fluoride complexes, Table

5.4, as well as in UO2(O2)2, Table 5.2. Like the U-Oyl and U-Operoxo bonds, the similarity of the

Page 197: Relativistic Quantum Chemistry Applied to Actinides

175

O-Operoxo bonds between the fluoro-peroxo and hydroxo-peroxo species results in similar

calculated frequencies for the O2 stretching vibrational modes.

The dihydroxo-diperoxo complex, UO2(OH)2(O2)24-

, like its fluoride analogue has two

conformers, a cis structure (C2v or butterfly arrangement of the peroxo ligands) and a trans

structure (D2h arrangement of the peroxo ligands). The trans structure was found to be more

stable than the cis conformer by 0.8 kcal/mol in aqueous solution. These structures are therefore

iso-energetic, in contradiction to the case in UO2(O2)22-

, where the cis orientation of the peroxo

ligands is favored by 3.1 kcal/mol. To explain the energy difference between the two conformers

of UO2(O2)22-

, we examined the nature of the σ(O2)-type orbitals and found some overlap across

the two peroxo groups in the C2v structure (see above and Figure 5.4). Examination of the low-

energy valence orbitals of cis- and trans- UO2(OH)2(O2)24-

reveals MO-30 to be of

predominantly σ(O-H) character, Figure 5.8. These orbitals however have significant σ(O2)-type

contributions. From an analysis of the orbital compositions, the σ(O2)-type contributions to MO-

30 are larger for the trans conformer of UO2(OH)2(O2)24-

. This compensates for the lacking

overlap between the two peroxo units, Figure 5.4, and is responsible for the iso-energetic nature

of the two conformers. Similar arguments are applicable to the fluoride analogues. As for

UO2F2(O2)24-

, attempts to optimize the structure of UO2(OH)2(O2)24-

in the gas phase failed as

the molecule fragmented into its component anionic pieces. Zehnder et al. have recently

characterized Na6[UO2(O2)2(OH)2](OH)2·14H2O using single crystal X-ray diffraction

techniques.70

They found that the UO2(OH)2(O2)24-

anion in Na6[UO2(O2)2(OH)2](OH)2·14H2O

has two trans peroxo groups and two trans hydroxo ligands. Compared to

Na6[UO2(O2)2(OH)2](OH)2·14H2O, (U-Oyl, 1.862 Å; U-Operoxo, 2.289/2.308 Å; U-OOH 2.388 Å

and O-Operoxo, 1.480 Å) the structural parameters of trans-UO2(OH)2(O2)24-

that we optimized in

Page 198: Relativistic Quantum Chemistry Applied to Actinides

176

solution appears to be sufficiently close (within 0.03 Å), Table 5.5. The only exceptions are the

U-OH bonds which are about 0.09 Å longer than those found in the sodium salt. To confirm the

stabilization of the UO2F2(O2)24-

and UO2(OH)2(O2)24-

species in solution, we checked that the

errors in the total polarization charges obtained for the PCM calculations, a reflection of the

portion of the density lying outside the cavity, were in all cases less than 0.05. We also examined

the volume of the cavity and the spatial extent of the virtual molecular orbitals in the complexes.

Uranyl carbonates and nitrates. The formation of UO2(O2)(CO3)24-

and UO2(O2)2(CO3)4-

,

from UO2(H2O)52+

by addition of HO2- and the carbonate ion follow the reaction

UO2(H2O)52+

+ nHO2- + 3-nCO3

2- → UO2(O2)n(CO3)3-n

4- + (3-n) H2O + nH3O

+.

Figure 5.8: MO-30 of cis- and trans- UO2(O2)2(OH)24-

. These orbitals are mostly of σ(O-H)

character with substantial contributions from the O-O bonds of the peroxide.

Page 199: Relativistic Quantum Chemistry Applied to Actinides

177

The reaction energies were calculated as -59.8 and 6.3 kcal/mol for UO2(O2)(CO3)24-

and

UO2(O2)2(CO3)4-

respectively at the B3LYP/B1 level in aqueous solvent. For the analogous

nitrate species, UO2(O2)(NO3)22-

and UO2(O2)2(NO3)3-

, the reaction energies are -16.7 and 20.4

kcal/mol respectively. These reaction energies suggest that the diperoxo species should not be

observed in solution. In a contemporaneous paper, Grenthe et al. have reported the speciation of

the uranyl-peroxide-carbonate system using a combination of potentiometric and NMR

measurements.71

UO2(O2)(CO3)24-

and UO2(O2)(CO3)2-

were the only species with one uranyl

group identified in their experiments. Most of the other species observed in their data contained

two uranyl groups. The absence of UO2(O2)2(CO3)4-

, in the experimental speciation data of

Grenthe et al., even at high H2O2 concentrations, is in good agreement with the calculated

reaction energies obtained in this work. The reaction energies for the formation of the nitrate-

peroxo complexes are higher than those of their carbonate counterparts, in line with the fact that

NO3- is a much weaker ligand for uranyl than CO3

2-. The carbonate and carbonate-peroxo

complexes are shown in Figure 5.9.

Structurally, the U-Oyl bonds are shorter in the nitrate complexes than in the counterpart

carbonate complexes, Tables 6 and 7. Comparison of the nitrato-peroxo and carbonato-peroxo

complexes to the fluoro-peroxo, Table 5.4, the hydroxo-peroxo, Table 5.5, and bare peroxo,

Table 5.2, complexes indicates that the U-Oyl bonds become increasingly weaker as the

interactions between the uranyl and equatorial ligands become stronger. The calculated U-Oyl

bond lengths in UO2(CO3)34-

and UO2(NO3)3- obtained in either the gaseous phase or in solution

are about 0.01-0.02 Å longer than the experimental values. It should be noted that the

experimental values were obtained from solid state X-ray or neutron diffraction studies of alkali

metal salts of UO2(CO3)34-

and UO2(NO3)3-.72-73

Assuming that this discrepancy between the

Page 200: Relativistic Quantum Chemistry Applied to Actinides

178

calculated and experimental U-Oyl bond lengths observed in the tricarbonate and trinitrate

complexes can be transferred to the peroxo complexes brings the calculated U-Oyl bond lengths

Figure 5.9: The structures of UO2(CO3)34-

and its peroxo derivatives optimized at the B3LYP/B1

level in aqueous solution. The analogous nitrate complexes possess similar structural

frameworks.

in UO2(O2)(CO3)24-

into agreement with the experimental value of 1.825-1.827 Å obtained from

the crystallographic data of K4[UO2(O2)(CO3)2].H2O.12

The U-Onitrate bond length in UO2(NO3)3- was calculated as 2.503 and 2.500 Å in the gas

and aqueous phases respectively, Table 5.7. In contrast, the solvent effect on the U-Ocarbonate bond

length is significantly higher, Table 5.6. Similar to the U-Oyl bonds, the U-Onitrate and U-Ocarbonate

bonds become weaker on coordination of a peroxo group. In the calculations carried out in

aqueous solution, the U-Ocarbonate bonds increase from 2.462 Å in the tricarbonate complex to

Page 201: Relativistic Quantum Chemistry Applied to Actinides

179

2.548 Å in the diperoxo-monocarbonate complex. The U-Onitrate bonds in the diperoxo-

mononitrate complex are about 2.714 Å long. The U-Ocarbonate bonds distal to the peroxo group in

UO2(O2)(CO3)24-

are slightly shorter (by 0.005 and 0.002 Å in the gaseous and aqueous phases

respectively) than those proximal to the peroxo group, Table 5.6. Similarly the distal U-Onitrate

bonds in UO2(O2)(NO3)22-

are shorter than their proximal counterparts by 0.047 and 0.030 Å in

the gas and aqueous phases respectively, Table 5.7. The trend in the U-Operoxo bond lengths in the

ternary peroxo-nitrate and peroxo-carbonate complexes is similar to those observed for the U-Oyl

and U-Ocarbonate/nitrate bonds.

In the calculations with the PCM solvation model, the bond between the oxygen atoms of

the peroxide ligands was calculated to be about 1.442 Å long in UO2(O2)22-

and 1.471 Å long in

UO2(O2)22-

, Table 5.2. This bond is slightly longer (about 1.460 Å) in the monoperoxo-fluoro

complexes, Table 5.4, as well as in the monoperoxo-carbonate complex (1.459 Å) and in the

monoperoxo-nitrate complex (1.451 Å). The calculated O-O bond length in UO2(O2)(CO3)24-

is

in good agreement with the experimental value of 1.469 Å obtained Goff et al.12

in their work on

K4[UO2(O2)(CO3)2].1H2O.

The calculated vibrational frequencies of the peroxo-carbonate and peroxo-nitrate

complexes are also presented in Table 5.6 and Table 5.7 respectively. The modes associated with

asymmetric and symmetric stretching of the uranyl group in UO2(CO3)34-

were calculated as 839

cm-1

and 773 cm-1

respectively in the gas phase. This calculated symmetric stretching frequency

is within the range obtained previously by Schlosser et al.32

and de Jong et al20

. The inclusion of

the solvent effects with the PCM model has little effect on the symmetric stretching frequency

but reduces the asymmetric stretching mode by about 28 cm-1

in comparison to the gas phase.

Page 202: Relativistic Quantum Chemistry Applied to Actinides

180

Table 5.6: Calculated structural properties and vibrational frequencies (cm-1

) of UO2(CO3)34-

and

its peroxo derivatives obtained at the B3LYP/B1 level in the gas phase and in aqueous solution.

Bond Lengths (Å) Uranyl Stretching

U-Oyl U-Ocarbonate U-Operoxo O-Operoxo Asymm. Symm.

UO2(CO3)34-

Gas phase 1.822 2.535 839 773

Solution 1.828 2.462 811 775

Experimental 1.81a 2.44

a 889

b 813

b

UO2(O2)(CO3)24-

Gas phase 1.848 2.636/2.641 2.272 1.461 796 730

Solution 1.856 2.498/2.500 2.255 1.459 761 739

Experimental 1.825-1.827b 2.429-2.473

b 2.238-2.255

b 1.469

b 766.5

b

UO2(O2)2(CO3)4-

Solution 1.883 2.548 2.286/2.296 1.475 708 706

a Reference

72

b Reference

74 c Reference

12

The calculated IR active asymmetric uranyl stretching frequency deviates from the experimental

value by about 71 cm-1

in the aqueous phase while its symmetric counterpart has an error of

about 31 cm-1

.74

In contrast, the agreement between the B3LYP/B1 model and the experimental

vibrational frequencies is improved for UO2(NO3)3-, Table 5.7. The symmetric stretching mode

was calculated to have a frequency of 870 cm-1

in good agreement with solid state Raman

spectroscopic measurements (876-886 cm-1

) of uranyl nitrate salts.75

The gas phase calculations

Page 203: Relativistic Quantum Chemistry Applied to Actinides

181

provided a value of 875 cm-1

for this vibrational mode, little change from the calculations with

the PCM model and in good agreement with the solid-state crystal measurements. The

frequencies of the uranyl asymmetric stretching vibrational mode were found to be 943.1 and

967.2 cm-1

in the potassium and ammonium crystalline salts from solid-state IR measurements.75

The calculated frequencies of this vibrational mode in the gaseous and aqueous phases are within

14-38 cm-1

of these experimental values, Table 5.7.

For the peroxo derivative of UO2(CO3)34-

and UO2(NO3)3-, the calculated frequencies of

the asymmetric and symmetric uranyl stretching modes decrease by 30-55 cm-1

for each

carbonate substitution in the peroxo-carbonate series, Table 5.6, and about 70-100 cm-1

for each

nitrate substitution in the peroxo-nitrate series, Table 5.7. For the mono-peroxo carbonate

complex, the calculated frequencies of the symmetric uranyl stretching mode, 739 cm-1

in

solution and 730 cm-1

in the gas phase, are in agreement with the experimental value of 766.5

cm-1

obtained during Raman spectral measurements of K4[UO2(O2)(CO3)2].H2O.12

Another

notable vibrational mode in UO2(O2)(CO3)24-

is the C-O stretch, calculated at 1054 cm-1

in good

agreement with the experimental value of 1053 cm-1

in the potassium hydrate complex.12

The

vibrational mode associated with the C-O3 bending of both carbonate groups (asymmetric out of

plane deformations) is Raman active and was calculated to have frequencies of respectively 847

and 842 cm-1

in the gaseous and aqueous phase calculations on UO2(O2)(CO3)24-

. The same

mode is also Raman active in the calculated spectra of UO2(CO3)34-

, 849 and 842 cm-1

in the

gaseous and aqueous phases respectively. Peaks corresponding to this vibrational mode were

observed between 849 and 879 cm-1

in the combined Raman and infra-red (IR) work of

Anderson et al.76

on K4[UO2(CO3)3]. In addition, de Jong et al.20

obtained a value of 845 cm-1

in

their theoretical study of uranyl carbonate using the local density approximation (LDA) with

Page 204: Relativistic Quantum Chemistry Applied to Actinides

182

Table 5.7: Calculated structural properties and vibrational frequencies (cm-1

) of UO2(NO3)3- and

its peroxo derivatives obtained at the B3LYP/B1 level in the gas phase and in aqueous solution.

Bond Lengths (Å) Uranyl Stretching

U-Oyl U-Onitrate U-Operoxo O-Operoxo Asymm. Symm.

UO2(NO3)3-

Gas phase 1.776 2.503 957 875

Solution 1.776 2.500 929 870

Experimental 1.77a 2.48-2.50

a 943-967

b 875-886

b

UO2(O2)(NO3)22-

Gas phase 1.821 2.662/2.709 2.187 1.446 858 787

Solution 1.822 2.589/2.619 2.198 1.451 824 777

UO2(O2)2(NO3)3-

Solution 1.869 2.714 2.252/2.264 1.473 735 703

a Reference

73 b Reference 75

diffuse basis sets. We note that the peak at 841.7 cm-1

in the Raman spectrum of

K4[UO2(O2)(CO3)2].2.5H2O was however labeled as the O-O symmetric stretching mode by Goff

et al.12

The frequency of the O-O symmetric stretching mode in UO2(O2)(CO3)24-

was however

calculated to be at 890 and 894 cm-1

in the gas phase and in solution respectively. This particular

mode was calculated to be at 944 cm-1

in PCM calculations on H2O2 in contrast to the

experimental value of 875 cm-1

. It is most likely the case that the assignment of the strong peak

at 841.7 cm-1

to the O-O symmetric stretching mode by Goff et al. is correct given the seemingly

large error in the calculated frequencies for this mode. The symmetric and asymmetric stretching

Page 205: Relativistic Quantum Chemistry Applied to Actinides

183

of the U-Operoxo bonds in UO2(O2)(CO3)24-

were calculated respectively as 383 and 375 cm-1

in

the gas phase and 394 and 387 cm-1

in solution. The asymmetric stretching modes, which were

calculated to possess low IR intensities, might account for the peak at 431.0 cm-1

in the Raman

spectrum of K4[UO2(O2)(CO3)2].2.5H2O.

Bond Orders in the uranyl and uranyl peroxo complexes. The population-based Mayer bond

orders obtained for all the bonds in the uranyl and uranyl peroxo complexes studied in this work

are collected in Table 5.8. The calculated bond orders for the U-Oyl bonds in UO22+

,

UO2(H2O)52+

, UO2(NO3)3-, UO2(OH)4

2-, UO2(CO3)3

4- and UO2F4

2- decrease down the series. The

orders for these bonds range from 2.53 in the bare dication to 2.29 in the tetrafluoro complex.

These values indicate that the U-Oyl bonds possess significant triple bond character with the

presence of equatorial ligands diminishing the triple bond nature. From the degree of

perturbation of the U-Oyl bonds, as seen in the bond order reduction, it appears that the aquo and

nitrate ligands have the weakest covalent interaction with the uranyl groups. This is supported by

the U-Owater and U-Onitrate bond orders of 0.48 and 0.46-0.49 respectively that are far smaller than

the U-Ocarbonate, U-Ffluoride and U-Ohydroxide bond orders respectively of 0.68, 1.02 and 1.07, Table

5.8. It is most likely that a competition between the U-Oyl and U-Xligand bonds for uranium 6d

atomic contributions result in the weakening of the U-Oyl bonds with increasing π-donating

abilities of the equatorial ligand. This correlates with the fact that the U-OH bonds have the

highest bond order of the equatorial U-Xligand bonds. As previously mentioned, it has recently

been shown that there is an increase in the actinide 6d contributions to the An-OH bonds and a

simultaneous decrease in the An 6d contributions to the An-Oyl bonds on progressing down the

AnO22+

to AnO2(OH)42-

(An = U, Np and Pu) series.58

Page 206: Relativistic Quantum Chemistry Applied to Actinides

184

Table 5.8. The calculated Mayer bond orders in various uranyl complexes and their peroxo derivatives obtained at the B3LYP/B2

level using structures optimized at the B3LYP/B1 level.

U-Oyl U-Operoxo O-Operoxo U-Xligand U-Oyl U-Operoxo O-Operoxo U-Xligand

UO22+

2.53 Halides

UO2(O2) 2.42 1.23 1.01 UO2F42-

2.29 1.02

C2v-UO2(O2)22-

2.31 1.05/1.08 0.98 UO2(O2)F22-

2.31 1.07 0.98 1.00

UO2(O2)34-

2.25 1.03 0.98 UO2(O2)F33-

2.26 1.05 0.99 0.96

Aquo UO2(O2)2F3-

2.27 1.02 0.97 0.95

UO2(H2O)52+

2.43 0.46-0.49 UO2(O2)2F24-

2.26 1.04 0.99 0.94

UO2(O2)(H2O)2 2.40 1.20 0.34 Hydroxides

UO2(O2)(H2O)3 2.39 1.18 1.00 0.34 UO2(OH)42-

2.32 1.07

UO2(O2)(H2O)4 2.39 1.17 1.01 0.30 UO2(O2)(OH)22-

2.32 1.08 0.98 1.06

UO2(O2)2(H2O)2-

2.31 1.06 0.98 0.24 UO2(O2)(OH)33-

2.29 1.06 0.98 1.00-

1.03

UO2(O2)2(H2O)22-

UO2(O2)2(OH)3-

2.28 1.02 0.97 1.02

Trans 2.30 1.00/1.04 0.98 0.25 UO2(O2)2(OH)24-

2.26 1.04 0.99 0.99

Page 207: Relativistic Quantum Chemistry Applied to Actinides

185

Cis 2.32 1.00/1.04 0.97 0.24 Nitrates

Carbonates UO2(NO3)3- 2.38 0.48

UO2(CO3)34-

2.31 0.68 UO2(O2)(NO3)22-

2.35 1.13 1.00 0.36

UO2(O2)(CO3)24-

2.29 1.05 0.99 0.64 UO2(O2)2(NO3)3-

2.28 1.02-

1.04

0.98 0.28

UO2(O2)2(CO3)4-

2.26 1.00-1.02 0.98 0.61

Page 208: Relativistic Quantum Chemistry Applied to Actinides

186

Going forward, we are interested in using more comprehensive charge and orbital decomposition

schemes to examine the interplay between covalent and electrostatic interactions in determining

the strength of the equatorial bonds in actinide complexes.

For the uranyl peroxo complexes, UO2(O2), UO2(O2)22-

and UO2(O2)34-

, the calculated

bond orders for the U-Oyl bonds decrease from 2.42 to 2.25, reminiscent of the case for the

fluoro, hydroxo, aquo, nitrate and carbonate complexes. The U-Operoxo bond orders also decrease

from 1.23 in the monoperoxo species to 1.03 in the triperoxo complex. The inclusion of other

ligands in the equatorial region of UO2(O2) leads to a slight decrease in the U-Oyl bond orders. A

similar case is observed for the U-Operoxo bonds, with the calculated bond order depending on the

binding strength of the other equatorial ligands. For the strongly binding carbonate, hydroxo and

fluoro equatorial ligands, the U-Operoxo bond orders are about 1.05-1.08 while for the weakly

binding aquo and nitrate ligands, the U-Operoxo bond orders are about 1.13-1.20. The bond orders

for the O-O bonds of the peroxide ligand remain within the range 0.97-1.01 regardless of the

number of peroxo groups in the complex as well as the nature of the other equatorial ligands.

Conclusions

The structural and electronic properties of various uranyl peroxo complexes have been

examined using scalar relativistic DFT calculations. The aqueous-phase structures of the peroxo

complexes were modeled with the PCM solvation model. The reaction energies for the formation

of the uranyl peroxo complexes from their parent uranyl complexes, the relative stabilities of the

various structures of the peroxo complexes as well as the role of the equatorial peroxo group on

the trans-cis transformation of the uranyl moiety were all examined in the gaseous and aqueous

phases.

Page 209: Relativistic Quantum Chemistry Applied to Actinides

187

The affinity of the peroxo ligand for the uranyl group far exceeds that of the F-, OH

-,

CO32-

, NO3- and H2O ligands. The reactions leading to the formation of the various uranyl-

monoperoxo complexes from UO2(H2O)52+

and HO2- were calculated to be significantly

exothermic in both the gaseous and aqueous phases. As a result, the U(VI) and peroxo

concentrations, both kinetic factors, are the major factors in experimentally identifying the

mononuclear uranyl-peroxo species studied in this work. It should be noted however that the

formation of the diperoxo UO2(O2)2X24-

/UO2(O2)2X4/3-

species from UO2(H2O)52+

and HO2- were

all calculated to be endothermic in aqueous medium. This implies that the monouranyl-diperoxo

complexes of the aquo, fluoro, hydroxo, carbonate and nitrate ligands would be absent in

solution in very good agreement with recent experimental data. On the other hand, attempts to

optimize the geometries of these complexes in the gas phase failed as they decomposed to the

component anions. This indicates the crucial roles of counter-ions in the crystallization of the

UO2(O2)2X24-

/UO2(O2)2X4/3-

species.

Examination of the electronic structures of the uranyl-peroxo complexes reveals that the

U-Operoxo bond is formed by overlap between U(VI) 5f orbitals and in-plane π antibonding

orbitals of the peroxo ligand. The σ and π bonding orbitals between the oxygen atoms of the

peroxo ligands are more stable than the orbitals of the uranyl moiety. There is however

significant π-π mixing between the orbitals of the peroxo ligand and the π(d) orbitals of the

uranyl. The importance of π(O2)- π(f) mixing is higher for the diperoxo and triperoxo complexes

as the inclusion of the second and third peroxo ligands further stabilizes the uranyl π(f) orbitals.

For UO2(O2)22-

, a cis arrangement of the peroxo groups was calculated to be more stable than the

D2h structure which features a trans arrangement of the equatorial peroxo groups. The origin of

this difference was found to be the presence of an overlap between the distal oxygen atoms of the

Page 210: Relativistic Quantum Chemistry Applied to Actinides

188

two peroxo groups in the σ(O2) orbitals. In contrast for the UO2(O2)2X24-

(for X = F- and OH

-)

species, the structures with cis and trans peroxo groups are iso-energetic as the σ(O2) orbitals

now contain overlap with the σ(O-H) and F2p orbitals.

The trends in the structures of the uranyl-peroxo complexes with the F-, OH

-, CO3

2-, NO3

-

and H2O ligands are similar to those previously observed in other uranyl complexes. Inclusion of

the peroxo ligand weakens the U-Oyl bonds resulting in sequentially decreasing uranyl

vibrational frequencies. The O-O bond of the peroxo complexes is mostly centered at 1.455-

1.480 Å and as such the O-O stretching vibrational mode is found between 840 and 940 cm-1

.

The calculated bond orders of the O-O bonds were found to be between 0.96-1.02, in good

correlation with the little influence on the O-O bond lengths by the type and nature of equatorial

ligands. The U-Operoxo bond-lengths are somewhat more sensitive to the type and number of

coordinated anionic ligands. The U-Operoxo stretching modes were calculated to be between 330-

419 cm-1

, with the symmetric mode being IR active and the asymmetric mode, Raman active.

References

1. Sigmon, G. E.; Ling, J.; Unruh, D. K.; Moore-Shay, L.; Ward, M.; Weaver, B.; Burns, P.

C., J. Am. Chem. Soc. 2009, 131, 16648.

2. Clarens, F.; de Pablo, J.; Casas, I.; Gimenez, J.; Rovira, M.; Merino, J.; Cera, E.; Bruno,

J.; Quinones, J.; Martinez-Esparza, A., J. Nucl. Mater. 2005, 345, 225.

3. Gimenez, J.; Clarens, F.; Casas, I.; Rovira, M.; de Pablo, J.; Bruno, J., J. Nucl. Mater.

2005, 345, 232.

4. Kubatko, K. A.; Forbes, T. Z.; Klingensmith, A. L.; Burns, P. C., Inorg. Chem. 2007, 46,

3657.

Page 211: Relativistic Quantum Chemistry Applied to Actinides

189

5. Ling, J.; Wallace, C. M.; Szymanowski, J. E. S.; Burns, P. C., Angew. Chem. Int. Edit.

2010, 49, 7271.

6. Sigmon, G. E.; Weaver, B.; Kubatko, K. A.; Burns, P. C., Inorg. Chem. 2009, 48, 10907.

7. Unruh, D. K.; Burtner, A.; Pressprich, L.; Sigmon, G. E.; Burns, P. C., Dalton T. 2010,

39, 5807.

8. Corbel, C.; Sattonnay, G.; Guilbert, S.; Garrido, F.; Barthe, M. F.; Jegou, C., J. Nucl.

Mater. 2006, 348, 1.

9. Armstrong, C. R.; Nyman, M.; Shvareva, T.; Sigmon, G. E.; Burns, P. C.; Navrotsky, A.,

P. Natl. Acad. Sci. USA 2012, 109, 1874.

10. Miro, P.; Pierrefixe, S.; Gicquel, M.; Gil, A.; Bo, C., J. Am. Chem. Soc. 2010, 132,

17787.

11. Vlaisavljevich, B.; Gagliardi, L.; Burns, P. C., J. Am. Chem. Soc. 2010, 132, 14503.

12. Goff, G. S.; Brodnax, L. F.; Cisneros, M. R.; Peper, S. M.; Field, S. E.; Scoft, B. L.;

Runde, W. H., Inorg. Chem. 2008, 47, 1984.

13. Meca, S.; Martinez-Torrents, A.; Marti, V.; Gimenez, J.; Casas, I.; de Pablo, J., Dalton T.

2011, 40, 7976.

14. Zanonato, P. L.; Di Bernardo, P.; Grenthe, I., Dalton T. 2012, 41, 3380.

15. Nyman, M.; Rodriguez, M. A.; Campana, C. F., Inorg. Chem. 2010, 49, 7748.

16. Austin, J. P.; Sundararajan, M.; Vincent, M. A.; Hillier, I. H., Dalton T. 2009, 5902.

17. Batista, E. R.; Martin, R. L.; Hay, P. J., J. Chem. Phys. 2004, 121, 11104.

18. Bühl, M.; Sieffert, N.; Chaumont, A.; Wipff, G., Inorg. Chem. 2011, 50, 299.

19. Bühl, M.; Sieffert, N.; Chaumont, A.; Wipff, G., Inorg. Chem. 2012, 51, 1943.

Page 212: Relativistic Quantum Chemistry Applied to Actinides

190

20. de Jong, W. A.; Apra, E.; Windus, T. L.; Nichols, J. A.; Harrison, R. J.; Gutowski, K. E.;

Dixon, D. A., J. Phys. Chem. A 2005, 109, 11568.

21. de Jong, W. A.; Harrison, R. J.; Nichols, J. A.; Dixon, D. A., Theor. Chem. Acc. 2001,

107, 22.

22. Garcia-Hernandez, M.; Lauterbach, C.; Krüger, S.; Matveev, A.; Rösch, N., J. Comput.

Chem. 2002, 23, 834.

23. Han, Y. K., J. Comput. Chem. 2001, 22, 2010.

24. Han, Y. K.; Hirao, K., J. Chem. Phys. 2000, 113, 7345.

25. Hay, P. J.; Martin, R. L., J. Chem. Phys. 1998, 109, 3875.

26. Kovacs, A.; Konings, R. J. M., J. Mol. Struc-Theochem. 2004, 684, 35.

27. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2010, 114, 1957.

28. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2011, 115, 14110.

29. Odoh, S. O.; Walker, S. M.; Meier, M.; Stetefeld, J.; Schreckenbach, G., Inorg. Chem.

2011, 50, 3141.

30. Pereira, C. C. L.; Marsden, C. J.; Marcalo, J.; Gibson, J. K., Phys. Chem. Chem. Phys.

2011, 13, 12940.

31. Privalov, T.; Schimmelpfennig, B.; Wahlgren, U.; Grenthe, I., J. Phys. Chem. A 2002,

106, 11277.

32. Schlosser, F.; Moskaleva, L. V.; Kremleva, A.; Krüger, S.; Rösch, N., Dalton T. 2010,

39, 5705.

33. Schreckenbach, G.; Hay, P. J.; Martin, R. L., J. Comput. Chem. 1999, 20, 70.

34. Schreckenbach, G.; Shamov, G. A., Acc. Chem. Res. 2010, 43, 19.

35. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109, 10961.

Page 213: Relativistic Quantum Chemistry Applied to Actinides

191

36. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2006, 110, 12072.

37. Shamov, G. A.; Schreckenbach, G.; Vo, T. N., Chem-Eur. J. 2007, 13, 4932.

38. Straka, M.; Kaupp, M., Chem. Phys. 2005, 311, 45.

39. Tsushima, S.; Uchida, Y.; Reich, T., Chem. Phys. Lett. 2002, 357, 73.

40. Vallet, V.; Macak, P.; Wahlgren, U.; Grenthe, I., Theor. Chem. Acc. 2006, 115, 145.

41. Vukovic, S.; Watson, L. A.; Kang, S. O.; Custelcean, R.; Hay, B. P., Inorg. Chem. 2012,

51, 3855.

42. Frisch, M. J.; et al Gaussian 03, Revision C.02. 2004.

43. Miertus, S.; Scrocco, E.; Tomasi, J., Chem. Phys. 1981, 55, 117.

44. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74, 1245.

45. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100, 7535.

46. Lim, I. S.; Stoll, H.; Schwerdtfeger, P., J. Chem. Phys. 2006, 124.

47. Becke, A. D., J. Chem. Phys. 1993, 98, 5648.

48. Lee, C. T.; Yang, W. T.; Parr, R. G., Phys. Rev. B. 1988, 37, 785.

49. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J., J. Phys. Chem. 1994, 98,

11623.

50. Laikov, D. N.; Ustynyuk, Y. A., Russ. Chem. B+ 2005, 54, 820.

51. Laikov, D. N., Chem. Phys. Lett. 2005, 416, 116.

52. Denning, R. G., Struct. Bond. 1992, 79, 215.

53. Denning, R. G., J. Phys. Chem. A 2007, 111, 4125.

54. Denning, R. G.; Foster, D. N. P.; Snellgrove, T. R.; Woodwark, D. R., Mol. Phys. 1979,

37, 1089.

Page 214: Relativistic Quantum Chemistry Applied to Actinides

192

55. Denning, R. G.; Green, J. C.; Hutchings, T. E.; Dallera, C.; Tagliaferri, A.; Giarda, K.;

Brookes, N. B.; Braicovich, L., J. Chem. Phys. 2002, 117, 8008.

56. Denning, R. G.; Snellgrove, T. R.; Woodwark, D. R., Mol. Phys. 1976, 32, 419.

57. Denning, R. G.; Snellgrove, T. R.; Woodwark, D. R., Mol. Phys. 1979, 37, 1109.

58. Odoh, S. O.; Reyes, J. A.; Schreckenbach, G., Inorg. Chem. 2012, submitted.

59. Ingram, K. I. M.; Haller, L. J. L.; Kaltsoyannis, N., Dalton T. 2006, 2403.

60. Allen, P. G.; Bucher, J. J.; Shuh, D. K.; Edelstein, N. M.; Reich, T., Inorg. Chem. 1997,

36, 4676.

61. Jones, L. H.; Penneman, R. A., J. Chem. Phys. 1953, 21, 542.

62. Toth, L. M.; Begun, G. M., J. Phys. Chem. 1981, 85, 547.

63. Tsushima, S.; Yang, T. X.; Suzuki, A., Chem. Phys. Lett. 2001, 334, 365.

64. Vallet, V.; Wahlgren, U.; Schimmelpfennig, B.; Moll, H.; Szabo, Z.; Grenthe, I., Inorg.

Chem. 2001, 40, 3516.

65. Garcia-Hernandez, M.; Willnauer, C.; Krüger, S.; Moskaleva, L. V.; Rosch, N., Inorg.

Chem. 2006, 45, 1356.

66. Clark, D. L.; Conradson, S. D.; Donohoe, R. J.; Keogh, D. W.; Morris, D. E.; Palmer, P.

D.; Rogers, R. D.; Tait, C. D., Inorg. Chem. 1999, 38, 1456.

67. Moll, H.; Reich, T.; Szabo, Z., Radiochim. Acta 2000, 88, 411.

68. Quiles, F.; Burneau, A., Vibrat. Spectrosc. 2000, 23, 231.

69. Quiles, F.; Chinh, N. T.; Carteret, C.; Humbert, B., Inorg. Chem. 2011, 50, 2811.

70. Zehnder, R. A.; Batista, E. R.; Scott, B. L.; Peper, S. M.; Goff, G. S.; Runde, W. H.,

Radiochim. Acta 2008, 96, 575.

Page 215: Relativistic Quantum Chemistry Applied to Actinides

193

71. Zanonato, P. L.; Di Bernardo, P.; Szabo, Z.; Grenthe, I., Dalton T. 2012, Advance

Article.

72. Amayri, S.; Reich, T.; Arnold, T.; Geipel, G.; Bernhard, G., J. Solid State Chem. 2005,

178, 567.

73. Barclay, G. A.; Sabine, T. M.; Taylor, J. C., Acta Crystallogr. 1965, 19, 205.

74. Allen, P. G.; Bucher, J. J.; Clark, D. L.; Edelstein, N. M.; Ekberg, S. A.; Gohdes, J. W.;

Hudson, E. A.; Kaltsoyannis, N.; Lukens, W. W.; Neu, M. P.; Palmer, P. D.; Reich, T.; Shuh, D.

K.; Tait, C. D.; Zwick, B. D., Inorg. Chem. 1995, 34, 4797.

75. McGlynn, S. P.; Neely, W. C.; Smith, J. K., J. Chem. Phys. 1961, 35, 105.

76. Anderson, A.; Chieh, C.; Irish, D. E.; Tong, J. P. K., Can. J. Chem. 1980, 58, 1651.

77. Hratchian, H. P.; Sonnenberg, J. L.; Hay, P. J.; Martin, R. L.; Bursten, B. E.; Schlegel, H.

B., J. Phys. Chem. A 2005, 109, 8579.

78. Schreckenbach, G.; Hay, P. J.; Martin, R. L., Inorg. Chem. 1998, 37, 4442.

79. Vaughn, A. E.; Barnes, C. L.; Duval, P. B., Angew. Chem. Int. Edit. 2007, 46, 6622.

80. Villiers, C.; Thuery, P.; Ephritikhine, M., Angew. Chem. Int. Edit. 2008, 47, 5892.

Page 216: Relativistic Quantum Chemistry Applied to Actinides

194

Preface to Chapter 6

This chapter is based on a manuscript published in the journal “Chemistry-A European Journal”.

The full citation of the paper is as follows:

Samuel O. Odoh, Pan Q. Jiang, Grigory A. Shamov, Feiyue Wang, Mostafa Fayek and

Georg Schreckenbach, “Theoretical Study of the Reduction of Uranium(VI) Aquo

Complexes on Titania Particles and by Alcohols”, Chem.-A Eur. J., 2012, 18, 7117-7127.

The possibility of using titania particles as a getter for radionuclides in mine tailings and at waste

sites has motivated us to study the adsorption of uranium complexes on both a periodic rutile and a

nanoparticulate anatase surface. The role of surface defects in inducing the reduction of adsorbed

species were also explored using density functional theory. The electronic structure and roles of

these surface defects were compared to the photochemical reduction of hexavalent uranium by

organic alcohols.

All the calculations in the published manuscript and compiled in this chapter were carried out by

Samuel O. Odoh. The manuscript was prepared together with the other authors.

Copyright permissions have been obtained from the publisher of this journal, Wiley, as well as

from the other authors. Pan Q. Jiang and Grigory Shamov were postdoctoral research fellows in

Professor Schreckenbach‟s group. Dr Mostafa Fayek and Dr Feiyue Wang are professors in the

Department of Geology and Department of Chemistry at University of Manitoba respectively.

Page 217: Relativistic Quantum Chemistry Applied to Actinides

195

Chapter 6: Theoretical Study of the Reduction of Uranium (VI) Aquo

Complexes on Titania Particles and by Alcohols

Abstract

To provide insights into the adsorption and photoreduction of uranium (VI) on TiO2, we

have studied the structural and electronic properties of uranium (VI) aquo complexes adsorbed

on stoichiometric and non-stoichiometric or „defected‟ TiO2 surfaces and nanoparticles. Plane

wave calculations with the pure PBE density functional and the PBE+U approach were used to

study U(VI) complexes on a periodic rutile (110) slab. In addition, a nanoparticulate Ti38O76

cluster was used to simulate anatase nanoparticles. The electronic structures of the adsorbed

U(VI) complexes indicate that the photoreduction process is a consequence of the photocatalytic

properties of TiO2. The reduction of the adsorbed complexes can only occur if the energy of the

incident photon exceeds the semiconductor band gap. The gap states induced by single or

neighboring hydrogen atoms and oxygen vacancies at the rutile (110) surface cannot reduce

adsorbed U(VI) complexes as the unoccupied 5f orbitals are found deeper in the conduction

band. In the absence of a solid substrate, photoreduction proceeds by abstraction of a hydrogen

atom from water or organic molecules present in solution. Photoreduction by chlorophenol

results in lower product yield than reduction by aliphatic alcohols. This is because the triplet

uranyl-chlorophenol complex is much more stable than similar complexes formed with

methanol and ethanol. In the case of water, the hydroxyl photoproduct easily re-oxidizes the

pentavalent species formed. In addition, it is easier for the triplet uranyl-water complex to

decompose to the photo-reactants.

Page 218: Relativistic Quantum Chemistry Applied to Actinides

196

Introduction

The retention by adsorption and reduction of uranium (VI) on geological surfaces in the

natural environment is of great importance to the nuclear industry. The safe disposal of high-

level nuclear waste (HLNW) and the design of containment structures at mines and

environmental remediation models can be improved with better understanding of the chemistry

of uranium adsorption and reduction in the natural environment. It is therefore important that

chemical processes involving the adsorption/desorption, reduction of U(VI) and the oxidation of

U(IV) in natural solid and aqueous systems are investigated. These processes have been the

subject of many scientific studies.1-14

In addition to surface adsorption, the reduction of the

soluble and environmentally mobile U(VI) to the less soluble and relatively immobile U(IV)

species represents a very potent avenue for uranium immobilization.

Experimentally, the reduction of U(VI) has been observed on abiotic solid systems.2-3, 8,

13, 15-16 The most widely studied of these involves the redox coupling of the Fe(II)/Fe(III) system

to that of U(VI)/U(V) with the U(V) produced undergoing disproportionation to give U(IV) and

U(VI). Liger and co-workers observed the reduction of U(VI) by Fe(II) adsorbed on colloidal

hematite at high pH.8 The reduction of U(VI) by Fe(II) adsorbed on corundum (Al2O3) has also

been observed.10

In the design of trappings for mine tailings or stored nuclear waste, it is

important to ensure that the trap or its products do not constitute mobile systems that can easily

leach into the environment as a contaminant. Thus, although the sulfides and oxides of iron as

well as zero-valent iron have been shown to effectively reduce U(VI), their practical use in

trappings may be limited due to the leaching of the oxidation byproducts such as sulfates and

sulfites into the ecosystem. Although the Fe(III) oxidation products are unreactive and immobile,

Page 219: Relativistic Quantum Chemistry Applied to Actinides

197

the by-products acidify the surrounding waters and remobilize other resident heavy metals such

as lead and cadmium.11-12

Roberts17

and Helean18-19

have suggested the use of titanium-based materials such as

titanate-containing ceramics as radionuclide getters because they are generally more resistant to

radiation damage and to geological alteration. The strong association of uranium with titanium

oxide (TiO2) in naturally uraniferous systems20

and kaolinite with titanium impurities21

allows us

to propose TiO2 as a novel uranium getter. In addition, uranium has been found to be intimately

associated as uraninite with anatase grains at the Nopal I uranium deposit in Mexico at a depth of

191 meters.20

Although TiO2 is more expensive, it has the advantages of greater adsorption

capacity, stability and relative environmental benignness in comparison to Al2O3, iron oxides and

iron pyrites. The reduction of U(VI) to U(IV) by rutile and anatase nanoparticles under dark

conditions has also been mentioned in very rare instances.2

Theoretical studies of the adsorption of the aquo-uranyl complex, [UO2(H2O)3]2+

and

other uranyl complexes22-24

at the rutile (110) and other mineral surfaces14, 25-28

have been

reported. Most of these theoretical reports have focused on elucidating the nature of the uranyl

adsorption (outer or inner-shell coordination) to the metal or metal oxide surface as well as

gauging the agreement between calculated and experimental geometrical parameters. The current

report summarizes a theoretical study of the electronic and structural properties of the uranyl

moiety adsorbed at the rutile (110) surface with a view to describing the mechanism for the

photoreduction of the adsorbed complexes. The electronic properties of uranyl complexes

adsorbed on non-stoichiometric titania systems with surface defects are also examined to

determine the relative energy positions of the unoccupied uranium 5f bands and the gap states

Page 220: Relativistic Quantum Chemistry Applied to Actinides

198

induced by the point defects. Finally, the photoreduction of U(VI) on TiO2 crystals is compared

to the quenching of the photoexcited U(VI) luminescent state by water and organic compounds.

The role of photoexcitation in the reductive immobilization of U(VI) on photocatalytic

titania surfaces is of interest.3, 13

A model involving the creation of an electron-hole (e-–h

+) pair

after photoexcitation has been used to describe the photocatalyzed reduction of CO2 and NO2 on

semiconducting metal oxide surfaces like rutile and anatase.29

Here, the photoexcited electron is

promoted into an orbital localized on the surface-adsorbed species. The highest occupied

molecular orbitals (HOMO) or valence band on TiO2 systems are dominated by contributions

from oxygen 2p atomic orbitals. Photoexcitation from the HOMO and subsequent electronic

relaxation, Figure 6.1a, could result in the creation of a hole at the top of the valence band and

the promotion of an electron into the uranium 5f orbital. Rapid disproportionation of the U(V)

thus formed could provide a pathway from this excited state to the formation of U(IV), R1-R5.

These reactions will be discussed in detail below.

R1: Surface adsorption: ≡TiO2 + [UO2(H2O)n]2+

→ [≡TiO2-UO2(H2O)n-2]2+

+ 2H2O

R2a: Photoexcitation: [≡TiO2-UO2(H2O)n-2]2+

+ hυ →[≡TiO2(+ h+

+1e-)-UO2(H2O)n-2]

2+*

R2b: Charge Transfer: [≡TiO2(+ h+

+ 1e-)-UO2(H2O)n-2]

2+*→ [≡TiO2(+ h

+)-UO2(+1e

-)(H2O)n-2]

2+*

R3: Electron-hole recombination: [≡TiO2(+ h+)-UO2(+ 1e

-)(H2O)n-2]

2+* → [≡TiO2-UO2(H2O)n-2]

2+

R4: U(V) disproportionation: 2U(V)(ads) → U(VI)(ads) + U(IV)(ads)

R5: Mixed oxide formation: UO2(ads) + 2UO3(ads) → U3O8(ads)

Page 221: Relativistic Quantum Chemistry Applied to Actinides

199

Figure 6.1: Possibilities for charge transfer to U(VI) complexes adsorbed on TiO2 crystals or

nanoparticles. a) Creation of an electron-hole (e-–h

+) pair after photoexcitation of a surface-

adsorbate complex. b) Excess electrons induced by surface or bulk defects are transferred to the

adsorbed U(VI) group initiating reduction.

Surface and bulk oxygen vacancies, single or neighboring surface hydroxyls and titanium

interstitials are commonly found on real TiO2 systems. They are crucial to the catalytic behavior

and chemistry of TiO2 crystals and surfaces.30-34

This is mainly due to the very high chemical

reactivities of these defects as a result of their excess electron(s). These defects and associated

unpaired electron states on TiO2 particles and surfaces have been extensively studied

Defect

1e- 1e-

Defect UO22+ + 2e-

+ UO22+

1a) 1b)

Photoexcitation

UO22+ + 1e-

1e

-

UO22+

Page 222: Relativistic Quantum Chemistry Applied to Actinides

200

theoretically and experimentally. Periodic density functional theory (DFT) calculations have

recently been used to show that the excess electrons from these defects can drive surface-

adsorbate charge transfer.31-32

The influence of the adsorbate‟s electronegativity on the charge

transfer from the point defects was also noted. From this perspective, it would be interesting to

determine the possibility of electron transfer from gap states on non-stoichiometric titania

systems to adsorbed U(VI) complexes, Figure 6.1b, as a possible pathway towards reductive

immobilization of uranium. The physical implication of this suggestion is the possibility of

reducing U(VI) by defected TiO2 crystals and surfaces under dark conditions (without

photocatalysis). The reduction of adsorbed U(VI) complexes by the unpaired electrons induced

by dopants, defects and impurities in TiO2 systems would be very important in the storage of

nuclear waste materials as the less mobile U(IV) in dark underground repositories.

In this work we employ relativistic density functional theory (DFT) calculations to

evaluate the suggestions posed above: 1) Can theoretical examination of the structural and

electronic properties of U(VI) complexes adsorbed on TiO2 systems provide insights into the

photocatalytic reduction process (Figure 6.1a)? (2) Can electron transfer from gap states induced

by single and neighboring surface hydroxyls result in the reduction of adsorbed uranyl

complexes (Figure 6.1b)? The first question relates to the behavior of TiO2 as a photocatalyst

while the second question concerns the ability of surface or sub-surface electrons induced by

defects, dopants or impurities to reduce U(VI) complexes adsorbed on TiO2 systems in the dark.

Finally, after elucidating the photoreduction mechanism on TiO2 systems, the quenching of

photoexcited U(VI) complexes by water and organic molecules in the absence of a

semiconductor substrate is also explored. The energetics of the hydrogen abstraction reaction as

well as the nature and stability of the triplet state uranyl-organic complexes are examined.

Page 223: Relativistic Quantum Chemistry Applied to Actinides

201

The current report is organized as follows: The first part of the discussion section is

focused on the description of the structural and electronic parameters of free U(VI) complexes as

well as stoichiometric and non-stoichiometric TiO2 surfaces and nanoparticles. Next, the

electronic and structural properties of U(VI) complexes adsorbed on stoichiometric TiO2 systems

are examined. The photoreduction of adsorbed uranyl complexes is then examined in light of the

electronic structures of the surface-adsorbate complexes. A periodic rutile (110) surface is

employed as a sample TiO2 surface while a Ti38O76 cluster is used to simulate a nanoparticle.

The quenching of photoexcited uranyl complexes by water and organic molecules are then

examined. Finally, the results of this work are summarized in the conclusions section.

Computational Details

Scalar relativistic periodic calculations were carried out with the VASP 5.235-38

package

using the projector augmented wave (PAW)39-40

approach with plane wave basis while

employing the PBE functional.41-42

To adequately localize and stabilize defect states using the

PBE+U approach, Hubbard correction terms, U, of 4.243

and 4.0 eV were used for titanium and

uranium atoms respectively.44

The U value of 4.2 eV has been shown to result in band gaps for

the rutile (110) and anatase (101) surfaces that agree well with experimental data in addition to

adequately reproducing the energetics of the electronic gap states induced by oxygen vacancies

on these surfaces.45

The spin-polarized calculations were carried out with a plane wave cutoff set

at 520 eV and a 3×3×1 k-point Monkhorst-Pack (MP) mesh was used. The electronic energy

convergence at each step and the geometry convergence criteria were stipulated as 10-4

eV and

0.01 eV Å-1

respectively. A rutile (110) slab with 192 atoms was cut from a bulk supercell

constructed from the optimized unit cell. Point defects were created by adding one or two

neighboring hydrogen atoms to the rutile (110) surface. The bottom layer in this slab was held

Page 224: Relativistic Quantum Chemistry Applied to Actinides

202

fixed at the bulk-optimized coordinates. Periodic slab images were separated by a vacuum of no

less than 12 Å to prevent interaction. Complexes belonging to the [UO2(OH)m(H2O)3-m]2-m

series

with the uranyl group bonded to two bridging surface oxygen atoms, Obb-, were used as the

U(VI) adsorbates in this work. As the adsorbates were added on only one side of the slab,

monopole, dipole and quadrupole corrections were carefully accounted for.

Anatase nanoparticles were represented by the Ti38O76 cluster model. This cluster is

about 12 Å × 16 Å in size and is therefore of nanoparticulate size. Gas phase geometry

optimizations were also carried out using the PBE and BP86 density functionals. Single point

calculations were then carried out with the B3LYP functional in aqueous solution using the

conductor-like screening model (COSMO).46-48

The atomic radii of the uranium, titanium,

oxygen and hydrogen atoms were chosen as 2.18 Å, 2.18 Å, 1.72 Å and 1.30 Å respectively. The

free uranyl complexes, water, methanol, ethanol, chlorophenol and all the complexes formed

between the uranyl and organic molecules were optimized in the gas phase and aqueous phases

with the PBE and B3LYP functionals. All atoms were described using all-electron basis sets of

triple-ζ polarized (TZP) quality while relativistic effects were included using the zeroth order

regular approximation (ZORA).49-51

All optimized geometries were characterized as local

minima on the potential energy surface using vibrational frequency analysis. The calculations on

the discrete clusters were carried out with the ADF suite of programs.52-53

Results and Discussion

Free Uranyl Complexes. The calculated structural and electronic properties of bare UO22+

and

the aquo and aquo-hydroxo complexes of the uranyl moiety are presented in Table 6.1.

[UO2(H2O)5]2+

is one of the dominant species in very acidic solutions while [UO2(H2O)3]2+

is the

Page 225: Relativistic Quantum Chemistry Applied to Actinides

203

Table 6.1: Calculated structural features and properties of U(VI) complexes obtained with the

PBE functional in the gas phase. The PBE functional and PBE+U approach (shown in italics)

were used with plane wave basis sets for the adsorbed complexes at the rutile (110) surface. The

U-OH2a average bond lengths are given.

U-Oyl (Å) U-OH2a (Å) υ UO2 (cm

-1) Gap (eV)

UO22+

1.724 963.2, 1058.8 2.286

[UO2(H2O)3]2+

1.762 2.381 909.1, 997.0 2.955

[UO2(H2O)5]2+

1.776 2.475 882.8, 970.0 2.921

[UO2(H2O)4(OH)]1+

1.804 2.569 837.4, 914.5 2.466

[UO2(H2O)3(OH)2] 1.817 2.484 819.6, 899.6 2.636

Adsorbed on a rutile (110) surface

U-Oyl (Å) U-OH2a (Å) U-Obb (Å) U-Tisurf (Å)

[UO2]2+

1.790 2.170 3.310

1.758 2.218 3.304

[UO2(H2O)3]2+

1.800 2.593 2.277/2.333 3.355

1.770 2.587 2.353/2.330 3.385

EXAFS (Ref. 29

) 1.78 2.46 2.31 3.02-3.56

[UO2(H2O)2(OH)]1+

1.815 2.557 2.510 3.493

Adsorbed on an anatase Ti38O76 cluster.

U-Oyl (Å) U-OH2a (Å) U-Obb / U-Tisurf (Å) 2 (cm

-1)

[UO2]2+

1.790 2.170/3.340 862.5, 933.2

[UO2(H2O)3]2+

1.800 2.572 2.281/3.405 846.9, 921.8

Page 226: Relativistic Quantum Chemistry Applied to Actinides

204

remaining motif after two equatorial ligands have been replaced by bridging oxygen atoms

belonging to the TiO2 surface or nanoparticle, Figure 6.2. The addition of aquo ligands to the

uranyl equatorial region results in a slight elongation of the U-Oyl bonds. The elongation of the

U-Oyl bonds results in the reduction of the symmetric and asymmetric vibrational stretching

frequencies of the uranyl cation, Table 6.1. The replacement of aquo ligands by hydroxo ligands

results in further elongation of the U-Oyl bonds and corresponding decrease in the uranyl

vibrational stretching frequencies.

Figure 6.2: Structure of UO22+

adsorbed at the rutile (110) surface at the bridging oxygen atoms.

The Osub, Obb and Tisurf, Hb and HHb atoms are labeled for easy identification. The oxygen,

titanium and uranium atoms correspond to the black, grey and white spheres respectively. The Hb

and HHb atoms are respectively the isolated and neighboring surface hydroxyl defects.

Page 227: Relativistic Quantum Chemistry Applied to Actinides

205

The lowest unoccupied orbitals of these U(VI) complexes are a set of empty 5f orbitals

while the HOMO in UO22+

is of σ-type character between the uranium and Oyl atoms. The

calculated HOMO-LUMO band gaps of the aquo complexes obtained with the PBE functional

are around 2.9 eV. The use of the B3LYP functional widens the gap between the occupied and

virtual orbitals. As an example, the calculated HOMO-LUMO gap of [UO2(H2O)5]2+

obtained

with the hybrid functional is 5.508 eV. Alignment of the MO energy diagrams of UO22+

(HOMO

and LUMO at -11.08 and -5.52 eV respectively) and [UO2(H2O)5]2+

(HOMO and LUMO at -

10.27 and -4.77 eV respectively) at the B3LYP level in aqueous solution reveals a slight shift of

the frontier orbitals to higher energies by about 0.8 eV in the pentaaquo complex. This form of

comparison is possible as they have the same total charge.

Clean Stoichiometric TiO2 Systems. Rutile surfaces and anatase nanoparticles were modeled

using periodic DFT calculations on a rutile (110) slab containing 192 atoms and molecular

calculations on a Ti38O76 cluster model respectively. The molecular cluster, which was cut from a

bulk anatase supercell is of nano-particulate size. We have used an anatase cluster rather than a

rutile cluster as the anatase phase is most stable for TiO2 nanoparticles below a size of 14 nm.54

The periodic calculations continue a line of theoretical and experimental investigations of the

adsorption of U(VI) complexes on the rutile (110) surface.22-24, 55

The (110) surface is known to

be the most stable compared to all other surfaces cut on crystalline rutile.56

This surface

constitutes about 56% of the exposed surfaces of crystalline rutile.57

The HOMO-LUMO gap for the optimized Ti38O76 cluster was calculated with the BP86

functional as 2.54 and 2.75 eV in the gas and aqueous phases respectively. These values are

smaller than the experimental value of 3.20 eV for bulk anatase.58

The first few highest occupied

frontier orbitals for this cluster were all found to possess predominantly O-2p atomic orbital

Page 228: Relativistic Quantum Chemistry Applied to Actinides

206

character while the lowest unoccupied MOs are mainly of Ti-3d character. The total and partial

electronic density of states (DOS) of a periodic rutile (110) slab optimized with the PBE

functional and the PBE+U approach using plane wave basis sets are presented in Figure 6.3.

Figure 6.3: Total and partial electronic density of states (DOS) obtained for a clean rutile (110)

slab while using with the PBE functional (left) and the PBE+U approach (right, U=4.2 eV).

Employing the U=4.2 eV correction increases the calculated band gap from 1.859 eV obtained

with the pure GGA functional to 2.418 eV.

These DOS show that the valence and conduction bands near the Fermi level are

dominated by the oxygen 2p and titanium 3d contributions respectively. In addition, the

electronic band gap was calculated as 1.859 with the PBE functional and 2.418 eV with the

PBE+U approach. These values are in good agreement with reports from other theoretical

investigations but significantly underestimate the experimental band gap of 3.03 eV for rutile.59-

Page 229: Relativistic Quantum Chemistry Applied to Actinides

207

60 The applied U-value of 4.2 eV appears sufficient for our needs as it has been shown to

adequately describe the energetics of the electronic gap states induced by surface defects.61

Non-Stoichiometric TiO2 Systems. The defect states induced by the presence of single (Hb)

and neighboring (HHb) surface hydroxyls on a periodic rutile (110) slab are shown in Figure 6.4.

The labeling of these surface hydroxyl defects are presented in Figure 6.2. The two excess

electrons in the HHb slab were calculated to be nearly degenerate in energy. The calculated

energies of the HHb gap states relative to bottom of the conduction band is in

Figure 6.4: Total and partial electronic density of states (DOS) obtained for hydroxylated rutile

(110) slabs with the PBE+U approach (U=4.2 eV for Ti 3d and U=4.0 eV for U 5f). The Ti-3d

gap states were calculated to be about 1 eV below the conduction band. The gap states caused by

two neighboring hydroxyl defects were calculated to be nearly degenerate in energy and are also

of Ti3+

character.

Page 230: Relativistic Quantum Chemistry Applied to Actinides

208

good agreement with the previous report of Bonapasta et al.62

The gap states are found at 1.097

and 0.906 eV below the conduction band for the slabs with Hb and HHb defects respectively.

They are therefore in the region observed from with photoelectron and electron energy loss

spectroscopic measurements.63-64

In addition, the gaps between the valence and conduction bands

on the hydroxylated slabs were calculated as around 2.191 and 2.141 eV for the Hb and HHb

slabs respectively, Figure 6.4. The band gap has been decreased by about 0.3-0.4 eV from the

value of 2.415 eV obtained for the clean stoichiometric surface. Finally, it is important to

mention that the excess unpaired electron(s) are localized on subsurface Ti atoms in both the Hb

and HHb slabs in essence turning these sites into Ti3+

centers.

U(VI) Complexes Adsorbed on Stoichiometric TiO2 Systems. The structural parameters of

[UO2]2+

and [UO2(H2O)3]2+

species adsorbed on a periodic rutile (110) slab are given in Table

6.1. These calculations were carried out with the PBE functional using plane wave basis sets.

The values obtained with the PBE+U approach are also included. The U-Oyl bond length in the

adsorbate complex, [UO2(H2O)2(OH)]1+

was calculated as 1.815 Å. This is longer than that

obtained for the adsorbed [UO2(H2O)3]2+

and [UO2]2+

complexes. It is however shorter than the

value of 1.843 Å obtained in the case of an adsorbed [UO2(OH)3]1-

group.55

In addition, there is

an increase in the U-Obb bond lengths (bonds between the uranium atom and the bridging surface

oxygen atoms, Obb, Figure 6.2) upon addition of an equatorial hydroxide ligand, Table 6.1. This

indicates a decrease in the surface adsorption strength with increasing pH. This decrease in

surface adsorption is confirmed by the increasing distances between the uranium atom and the

surface titanium atoms, U-Tisurf, Table 6.1. The increase in the U-Oyl and U-Obb bond lengths

upon addition of a hydroxide group is similar to the situation in the free complexes. This is seen

as a result of the uranyl ion forming stronger complexes with hydroxide and carbonate anions

Page 231: Relativistic Quantum Chemistry Applied to Actinides

209

than with water and the bridging oxygen atoms at the rutile surface.55

The use of the PBE+U

results in reduction of the calculated U-Oyl and U-Obb bond-lengths. The calculated U-Oyl and U-

Obb bond lengths obtained for the complex [UO2(H2O)3]2+

using the PBE+U approach are in

good agreement with EXAFS results65

, Table 6.1.

The bonds between the bridging oxygen atoms and the hexa-coordinated Ti atoms

beneath them at the clean rutile (110) surface, Obb-Tisurf, were calculated to be between 1.858 and

1.860 Å when the PBE+U approach was employed, Figure 6.2, in comparison with the

experimental value of 1.85 Å.66

The adsorption of UO22+

and [UO2(H2O)3]2+

results in significant

elongation of these surface bonds to about 1.994-2.027 and 1.948-1.973 Å respectively. In

addition, the bonds between the Ti atom directly beneath the adsorbed uranyl complex and the

tri-coordinated sub-surface oxygen atoms, Tisurf-Osub, were calculated as 1.908 and 1.936 Å for

the UO22+

and [UO2(H2O)3]2+

adsorbates respectively, Figure 6.2. These are much shorter than

was obtained from calculations, 2.060 Å and experimental work, 2.08 Å on the clean surface.66

It should also be noted that complexation of the bare uranyl or pentaaquo complex to the

TiO2 surface or nanoparticle results in slight elongation of the U-Oyl bonds by 0.01-0.07 Å. This

is in agreement with the results of EXAFS studies, Table 6.1.65

The elongation of the U-Oyl bond

lengths upon surface adsorption is reminiscent of a similar effect when equatorial aquo ligands

are replaced by hydroxide ligands, as has been discussed above. The effect of this is a reduction

in the calculated stretching modes of the uranyl group. The calculated symmetric and asymmetric

vibrational modes of the UO22+

and [UO2(H2O)3] 2+

groups adsorbed on the anatase nanocluster

are much lower than was obtained for the similar free complexes, Table 6.1. Experimental

confirmation (using Fourier transform infra-red spectroscopy, FTIR) of this change in the uranyl

vibrational wavenumbers has been recently reported.67

Page 232: Relativistic Quantum Chemistry Applied to Actinides

210

The electronic DOS for the surface-adsorbate complexes with UO22+

and [UO2(H2O)3]2+

adsorbed on the bridging oxygen atoms of a periodic rutile (110) slab are presented in Figure

6.5. Overall, the DOS for the surface-adsorbate complex is similar to that of the clean (110) slab

and dominated by contributions from the rutile slab. This is most likely a result of the low

uranyl surface coverages used in this work. The electronic band gap of the surface-adsorbate

complex with a [UO2]2+

adsorbate was calculated as 1.442 and 2.060 eV with the PBE

functional and PBE+U approach respectively. These are about 0.417 and 0.352 eV respectively

Figure 6.5: Total and partial electronic density of states (DOS) obtained for rutile (110) slabs

with an adsorbed UO22+

group obtained with the PBE+U approach (U=4.2 eV for Ti-3d and

U=4.0 eV for U-5f). The empty U-5f states were calculated to be just beneath the Ti-3d

dominated conduction band.

lower than those obtained for the clean (110) surface respectively. The U value of 4.0 eV has

been shown to be sufficient in describing the unoccupied 5f orbitals of uranium fuels and

Page 233: Relativistic Quantum Chemistry Applied to Actinides

211

clusters.68-69

Given that the experimental band gap of rutile is 3.0 eV, the results of the PBE+U

calculations suggest that the empty 5f bands can be expected to be found between 0 and 1 eV

beneath the Ti-3d conduction band. At the current level however a 5f band is found just beneath

the conduction band.

For the periodic slab with a [UO2(H2O)3]2+

adsorbate, the empty 5f band is shifted to

slightly higher energies and is now buried in the Ti-3d conduction band, Figure 6.6. The

electronic band gap of the rutile-[UO2(H2O)3]2+

surface adsorbate complex was calculated to be

2.136 eV. A look at the calculated DOS however suggests that given accurate replication of the

experimental rutile band gap, the 5f band should be just below (-0.2 to 0.8 eV) the Ti-3d

Figure 6.6: Total and partial electronic density of states (DOS) obtained for stoichiometric (left)

and surface hydroxyl defected (right) rutile (110) slabs with adsorbed [UO2(H2O)3]2+

obtained

with the PBE+U approach (U=4.2 eV for Ti-3d and U=4.0 eV for U-5f). The gap state induced

by the surface hydroxyl lies at lower energies than the empty U-5f bands.

Page 234: Relativistic Quantum Chemistry Applied to Actinides

212

conduction band. Examination of the electronic structure of the Ti38O76-[UO2(H2O)3]2+

and

Ti38O76-[UO2]2+

complexes also reveals the presences of empty U-5f states between the O-2p

and Ti-3d bands of the anatase cluster. Figure 6.7. In agreement with the PBE+U calculations on

the periodic slab, the empty U-5f band is located between 1 and 0.0 eV of the Ti-3d band. The

ordering and general description of the bands dominated by the O-2p, U-5f and Ti-3d remains

the same for the surface-adsorbate complexes of UO22+

and [UO2(H2O)3]2+

.

Figure 6.7: Electronic energy levels and frontier molecular orbitals obtained at the BP86/TZP

level for the a Ti38076-[UO2(H2O)3]2+

surface-adsorbate complex. The empty U-5f orbitals are

between 1.0 and 0.0 eV of the onset of the Ti-3d virtual orbitals.

Page 235: Relativistic Quantum Chemistry Applied to Actinides

213

The empty 5f states might be occupied during photoexcitation from the O-2p dominated

valence band. In this event, the adsorbed uranyl group acquires a pentavalent, U(V) character

with the simultaneous creation of a hole in the titania system. Indeed, examination of the lowest

triplet excited states of the Ti38O76-[UO2]2+

and Ti38O76-[UO2(H2O)3]2+

clusters reveals that an

electron is invariably excited from an O-2p orbital localized in the TiO2 cluster into a uranium 5f

orbital. For example, the calculated Mulliken spin densities on the Ti38O76 and [UO2(H2O)3]2+

groups in the lowest triplet state of Ti38O76-[UO2(H2O)3]2+

are 1.014 and 0.986 respectively.

Indeed TDDFT calculations of the first 50 excitations of the Ti38O76-[UO2(H2O)3]2+

cluster

reveal the final orbitals to be of distinctly U-5f character. The trapping of the excited electron in

the actinide 5f orbitals is in contrast to the localization of the electron-hole pair on the subsurface

atoms of a photoexcited rutile (110)-water surface adsorbate complex containing associatively

adsorbed water.70

Electronic transitions from the ground state to the triplet excited states are

forbidden however due to spin selection rules. This might indicate low probabilities for the

photoreduction pathway initiated by direct excitation into the lowest triplet states.

Conversely, initial photoexcitation into the Ti-3d dominated virtual orbitals (LUMO+5

and above) or the conduction band for the periodic slab may also be followed by fast electronic

decay and charge transfer into the 5f-band/orbitals of the adsorbed U(VI) species. This indirect

process of excitation of the valence electrons into the TiO2 conduction band followed by charge

transfer into the uranium-5f atomic orbitals is likely to be more plausible given the spatial

overlap of the MOs involved in the initial excitation, Figure 6.7. The nature of this charge-

transfer mechanism for the photoreduction of the adsorbed uranyl complexes indicates that the

incident energy must be equal or greater than the semiconductor band gap. This is in accordance

with the experimental work of Amadelli et al.1 It also indicates that the role of the TiO2 surface

Page 236: Relativistic Quantum Chemistry Applied to Actinides

214

or nanoparticle is mainly photocatalytic. A sequence of reactions, R1-R5 occurring between the

formation of the surface-adsorbate complex to the formation of mixed U(VI)/U(IV) oxides can

be posited. It should be noted that the direct photoexcitation (O-2p → U-5f) and indirect charge

transfer (O-2p → Ti-3d → U-5f) processes both involve excitation of electrons from the TiO2

valence band. In the case of the latter, the yield of the U(IV) formed would match the absorption

spectrum of the TiO2 system. This was also found to be the case by Amadelli et al.1

Concerning reactions R4 and R5, the presence of adsorbed polymeric uranyl species and

hole-scavengers will appear to favor the disproportionation process, R4 as these species afford a

polynuclear framework allowing disproportionation of the pentavalent intermediates. The

oxidation of the U(V) species by dissolved molecular oxygen is known to be relatively slow and

the precipitation of the adsorbed U(IV) would drive R4 forward under sustained irradiation.4, 71-72

The suspension of photo-irradiation results in the re-oxidation of U(IV) species by molecular

oxygen. This could lead to the formation of U(IV)/U(VI) mixed oxides, R5.1, 3, 13

U(VI) Complexes Adsorbed on Non-Stoichiometric TiO2. The possible reduction of U(VI)

complexes adsorbed on defected surfaces, Figure 6.1b, is similar to the photocatalyzed reduction

process on stoichiometric TiO2 slabs, Figure 6.1a. The only difference being that the U(VI)

complex would be reduced by the transfer of an electron from a deep-lying 3d1 electronic state

formed by the point defects on adsorption rather than by a photoexcited electron.

The electronic DOS of periodic rutile (110) slabs with an adsorbed [UO2(H2O)3]2+

group

and a surface hydroxyl defect obtained using the PBE+U approach is presented in Figure 6.6.

The uranyl and surface hydroxyl groups were placed on adjacent Obb bridging rows. In the

absence of the equatorial aquo ligands, the excess electron from the single hydroxyl defect is

Page 237: Relativistic Quantum Chemistry Applied to Actinides

215

localized on a Ti atom in the rutile slab in the most stable electronic state. Although this result

might suggest that the unpaired electron is not transferred to the adsorbed U(VI) complex, it has

been shown that the gap states predicted by the PBE+U approach are not as deeply localized as

those predicted by hybrid functionals.31, 34

Enforcing the localization of the unpaired electron on

the uranium atom of the adsorbed uranyl group increases the total energy by about 0.11 eV. This

is done by using the MAGMOM keyword in the VASP software package. The introduction of

three aquo ligands on the uranyl group, [UO2(H2O)3]2+

, Figure 6.8, further increases the energy

barrier to localization of the excess electrons on the uranium atom. A similar situation was

observed when the surface hydroxyl group was replaced by either neighboring hydroxyls or an

oxygen vacancy.

Figure 6.8: Calculated spin distributions of a rutile (110) slab with a surface-adsorbed

[UO2(H2O)3]2+

moiety and a surface hydroxyl defect. These calculations were carried out with

the PBE+U approach. Two side views of each slab are presented [Titanium: grey, Oxygen: red,

Uranium: blue and Hydrogen: pink].

Page 238: Relativistic Quantum Chemistry Applied to Actinides

216

Going back to the anatase molecular clusters, the MO energy level diagram of the

defected Ti38O76H (surface hydroxyl) cluster with an adsorbed [UO2(H2O)3]2+

is presented in

Figure 6.9. The delocalization of the excess electron state on several titanium atoms with a small

portion on the uranyl moiety is most likely an artifact of the geometry optimization with the

GGA functional.34

The gap state induced by the surface defect is found below the virtual U-5f

dominated orbitals. This state possesses appreciable U-5f atomic orbital contributions. The U-5f

contribution to the gap state was calculated to be about 28% with the BP86 functional but was

decreased to around 12% after a the single point calculation with the B3LYP functional. This

results in the reduction of the spin density on the adsorbed uranyl group by about 50%.

Figure 6.9: Electronic energy levels and MOs of a non-stoichiometric cluster (with a surface

hydroxyl)-adsorbate complex. The defect state induced by the hydroxyl is predominantly of Ti-

3d character. The Mulliken spin density on the uranyl complex is reduced by 50 % from 0.30 to

0.15 after a single point calculation with the B3LYP functional.

Page 239: Relativistic Quantum Chemistry Applied to Actinides

217

The combination of the results of the calculations on the periodic slab and the

nanoparticulate cluster model indicates that surface hydroxyl and oxygen vacancy defects play

little role in the reduction of adsorbed U(VI) complexes. The gap states induced by these defects

occur at lower energies than the empty U-5f bands. The addition of aquo ligands to the equatorial

region of the adsorbed uranyl further shifts the empty U-5f band away from the band gap. It

should also be noted that these defect species have tremendous affinities for adsorbed or

dissolved oxygen suggesting their role in U(VI) reduction would be even further limited under

oxic conditions.

Quenching of Photoexcited U(VI) Complexes by Water and Organic Molecules. As

discussed thus far, the role of TiO2 in the reductive immobilization of U(VI) complexes appears

to be mainly photocatalytic while surface defects play little or no role in the reduction of surface

adsorbed U(VI) complexes. In contrast, the photoreduction of U(VI) complexes by organic

molecules in the absence of a solid photocatalytic surface is well known from experimental

studies.5, 9, 72-74

Theoretical studies directed at understanding the quenching of photoexcited

actinyl dications, [AnO2]2+

by water have been done.74-77

A more generic reaction involving the

abstraction of a hydrogen atom from water and organic molecules in quenching the photoexcited

pentaaquo uranyl complex can be written as R6.

R6: Hydrogen atom abstraction: [UO2(H2O)n]2+,*

+ ROH → [UO2H(H2O)n]2+

+ RO.

Thermochemically, the quenching of the lowest triplet excited state of [UO2(H2O)5]2+

by

abstracting an hydrogen atom from water is significantly endoergic in comparison to similar

processes by aliphatic alcohols or 4-chlorophenol, Table 6.2. Photoreduction via abstraction of

the α-hydrogen atoms of methanol is more exoergic than abstraction from its alcohol-hydrogen

Page 240: Relativistic Quantum Chemistry Applied to Actinides

218

atom by about 6.6 to 8.1 kcal/mol. A similar preference for the α-hydrogen atom of ethanol was

also observed. The calculated reaction energies for R6 actually reflect the relative stability of the

radicals formed by either abstraction of the α-, β- and OH hydrogen atoms and is in very good

agreement with the experimental observation that the lowest triplet state of U(VI)9, 73

like

carbonyl radicals78

tend to preferentially abstract the α-hydrogen atoms of aliphatic alcohols.

Accounting for solvent polarity by performing single point calculations in methanol using

optimized gas phase structures did not alter the conclusions that can be drawn from the results

presented in Table 6.2. In the case of 4-chlorophenol, it appears that abstraction of the phenolic

hydrogen atom represents the only feasible photoreduction pathway. Abstraction of hydrogen

atoms from the 2- and 3- position of the phenyl ring were calculated to be endoergic at the

PBE/TZP level. In addition, the results of the DFT calculations suggest that the photooxidation

of chlorophenol is energetically more favored than quenching by the aliphatic alcohols or water.

It should however be noted that these thermodynamic calculations on R6 do not involve a

description of the interaction between the triplet excited state [UO2(H2O)5]2+

complex and the

ROH (quenching) molecules.

The nature of the complexes formed between the triplet excited state of [UO2(H2O)5]2+

and the quenching molecules (water, aliphatic alcohols or phenols) can be probed using

unrestricted DFT calculations. Examination of the calculated S(S+1) values for these uranyl-

quencher complexes indicates little or no spin contamination. The calculated lengths of the U-Oyl

bonds and the bonds between the uranyl oxo atom and the hydrogen atom to be abstracted in the

uranyl-quencher complex, Ooxo-Habs are presented in Table 6.2. Attempts to optimize the

geometries of the triplet state uranyl-quencher complex formed through the 2- and 3- hydrogen

atoms of chlorophenol and the β-hydrogen atoms of ethanol were unsuccessful. The instability of

Page 241: Relativistic Quantum Chemistry Applied to Actinides

219

Table 6.2: Calculated reaction energies (kcal/mol) obtained for R6 in aqueous solution obtained

at the PBE/TZP and B3LYP/TZP levels in addition to the calculated structural and electronic

properties of the triplet state uranyl-quencher complexes.

Quencher R6 Reaction Energies Uranyl-Quencher Complexesa

H atom PBE B3LYP U-Oylb Oyl-Habstracted

b Spin

c

Water OH 14.25 2.24 1.809/1.898 2.357/2.361 0.411

Methanol

α-H -12.20 -20.33 1.803/1.996 1.019 0.959

OH -4.13 -13.52 1.806/1.969 1.025 1.002

Ethanol

α-H -14.59 -22.36 1.808/1.966 1.041 0.947

β-H -4.54 -13.22

OH -4.81 -13.63 1.808/1.956 1.034 1.001

4-chlorophenol

2-H 5.62 -2.40

3-H 5.89 -1.95

OH -22.31 -31.28 1.822/1.882 1.430 1.001

a The structural features and electronic properties presented are from triplet uranyl-quencher

complexes whose geometries were optimized in aqueous solution at the B3LYP/TZP level. b The

longer U-Oyl bonds belong to the oxo atom involved in hydrogen abstraction. c This denotes the

partial spin localized on the radical formed after H abstraction from the quenching molecule.

the triplet state uranyl-quencher complexes formed via the hydrogen atoms at the 2- and 3-

phenyl positions reinforces the thermochemical calculations which indicated that the

Page 242: Relativistic Quantum Chemistry Applied to Actinides

220

photoreduction of U(VI) by chlorophenol can only proceed via the formation of the

chlorophenoxy radical.

In all the optimized triplet state uranyl-quencher complexes considered, there is complete

transfer of an electron to the uranyl group with the uranium atom being formally pentavalent,

Table 6.2. The distance between the uranyl oxo atom and the hydrogen atom being abstracted

from the quenching molecule, Ooxo-Habs, was calculated to be about 2.36 Å in the triplet

[UO2(H2O)5][H2O]2+,

* complex. This distance is rather large in comparison to the distances of

about 1.0 Å in the complexes formed with the aliphatic alcohols. The distance in the

[UO2(H2O)5][chlorophenol]2+,

* complex is about 0.5 Å longer than observed for the aliphatic

alcohols. In addition, the triplet state uranyl-quencher complex formed with chlorophenol was

calculated to be significantly relatively more stable (relative to the photo-products) than those

formed with water, methanol or ethanol, Figure 6.10. The longer Ooxo-Habs distance as well as the

greater stability of the triplet uranyl-chlorophenol complex is most likely due to stabilization via

electron delocalization from the aromatic ring.

The yield of any reduction products (e.g. U(IV) precipitate, alkene and alkyl radical

oxidation products) depends on the susceptibility of the ability of the triplet uranyl-quencher

complex to decomposition. If the triplet uranyl-quencher complex is „too‟ stable relative to

photoreduction products, decomposition subsequent to hydrogen abstraction will be more

endoergic. This will result in low yields for U(V) species, U(IV) formed on disproportionation of

the pentavalent complex, alkoxyl or hydroxyl radical and ketones formed from the alkoxyl

radicals. This is most likely the case for the photoreduction by chlorophenol, Figure 6.10, and is

in some agreement with the experimental work of Sarakha et al.72

They observed low product

yields even while obtaining very high charge transfer rates between the chlorophenol molecule

Page 243: Relativistic Quantum Chemistry Applied to Actinides

221

and the excited uranyl complex. This correlates well with the significant spin/charge transfer in

the triplet uranyl-chlorophenol complex, Table 6.2, and its high stability relative to the photo-

reactants and photo-products, Figure 6.10. In contrast, the triplet uranyl-water complex was

found to be the least stable with respect to the photo-reactants and the most stable relative to the

photo-products, Figure 6.10. This indicates that the uranyl-water triplet state complex would

most likely decompose back to the photo-reactants. This is in stark contrast to the other triplet

state complexes which would progress to the photo-products given the relative energies of the

forward and backward reactions. It should however be noted that we have only examined the

quenching of the lowest triplet excited state of the uranyl complex. Higher electronic excited

states might be involved in the photooxidation of water by uranyl complexes.

Figure 6.10: The quenching of the lowest triplet excited state of [UO2(H2O)5]2+

by water and

organic alcohols. The relative energies (ΔG, calculated at the BP86/TZVP level in aqueous

Page 244: Relativistic Quantum Chemistry Applied to Actinides

222

solution) of the reactants and products of reaction R6 are given with respect to the triplet uranyl-

quencher complex.

Conclusions

We have carried out a study of the structural and electronic properties of U(VI)

complexes adsorbed on TiO2 surfaces to obtain insights into the photoreduction mechanism of

uranium compounds on such surfaces. The possibility of charge transfer from surface defects

such as hydroxyls and oxygen vacancies to the adsorbed U(VI) complexes was explored as well.

The quenching of the lowest triplet excited state of U(VI) complexes by water and small organic

molecules was also elucidated for comparison to the photocatalytic role of the TiO2 surface. All

these calculations were carried out using DFT either on periodic extended systems with plane

waves or on cluster models with all-electron basis sets.

The structures of rutile surfaces as well as the structures of the adsorbed uranyl

complexes are well reproduced both in terms of absolute values and in structural trends by using

the PBE+U method in the periodic calculations. When compared with available experimental

data, the cluster model calculations with the PBE functional provide accurate descriptions of the

IR spectra of the uranyl complexes adsorbed on nanoparticles. Electronically, the empty 5f states

of the adsorbed uranyl complexes were calculated to be just below the Ti-3d conduction band.

The selection rules indicate that a charge-transfer mechanism which is O-2p → Ti-3d → U-5f is

most likely responsible for the photoreduction of U(VI) on TiO2 surfaces. This implies that the

photoreduction process on TiO2 surfaces and nanoparticles is photocatalytic in nature and does

not involve direct excitation of the uranyl complex. For such surface catalyzed reduction

processes to occur, the energy of the incident light must be equal to or exceed the semiconductor

band gap. The implication of this constraint is that the photoreduction yield must match the

Page 245: Relativistic Quantum Chemistry Applied to Actinides

223

absorption spectrum of the semiconductor. This is in agreement with previous experimental

reports further demonstrating the potency of the approach employed.

According to the computations, electron transfer from surface oxygen vacancies and

hydroxyl defects do not appear to be able to induce U(VI) reduction under dark conditions. This

is because the gap states induced by these defects are found nearer the valence band than the

empty 5f bands. The equatorial aquo ligands of the uranyl complex further shift the empty U-5f

band into the conduction band. This decreases the possibility of localizing the excess electrons

on the uranium atom.

In contrast to photoreduction on TiO2, the photoreduction by water, aliphatic alcohols and

chlorophenol is initiated by initial photoexcitation of the uranyl complex. The calculated reaction

energies indicate that, the hydroxyl radical formed by abstracting an hydrogen atom from a water

molecule is able to rapidly re-oxidize the U(V) species formed. The aliphatic alcohols and 4-

chlorophenol more easily quench the photoexcited uranyl complex by donating their α- and

phenolic hydrogen atoms respectively. The relative stabilities of the triplet state uranyl-quencher

complexes with respect to the pentavalent and alkoxyl (or phenoxyl) photoreduction products

appear crucial to the photo-product yield observed. The uranyl-chlorophenol triplet state

complex is more stable relative to the photo-products than similar complexes formed with the

aliphatic alcohols. This correlates well with previous experimental work in which very high

charge transfer quantum yields with very low yields for the photo-products were obtained for the

uranyl-chlorophenol system.

Page 246: Relativistic Quantum Chemistry Applied to Actinides

224

References

1. Amadelli, R.; Maldotti, A.; Sostero, S.; Carassiti, V., J. Chem. Soc. Faraday T. 1991, 87

(19), 3267-3273.

2. Bonato, M.; Allen, G. C.; Scott, T. B., Micro Nano. Lett. 2008, 3 (2), 57-61.

3. Eliet, V.; Bidoglio, G., Environ. Sci. Technol. 1998, 32 (20), 3155-3161.

4. Formosinho, S. J.; Burrows, H. D.; Miguel, M. D.; Azenha, M.; Saraiva, I. M.; Ribeiro,

A.; Khudyakov, I. V.; Gasanov, R. G.; Bolte, M.; Sarakha, M., Photoch. Photobio. Sci. 2003, 2

(5), 569-575.

5. Formosinho, S. J.; Miguel, M. D. M.; Burrows, H. D., J. Chem. Soc. Faraday T. I 1984,

80, 1717-1733.

6. Harazono, T.; Sato, S.; Fukutomi, H., B. Chem. Soc. Jpn. 1984, 57 (3), 768-770.

7. Lefevre, G.; Kneppers, J.; Fedoroff, M., J. Colloid Interf. Sci. 2008, 327 (1), 15-20.

8. Liger, E.; Charlet, L.; Van Cappellen, P., Geochim. Cosmochim. Ac. 1999, 63 (19-20),

2939-2955.

9. Nagaishi, R.; Katsumura, Y.; Ishigure, K.; Aoyagi, H.; Yoshida, Z.; Kimura, T., J.

Photoch. Photobio. A. 1996, 96 (1-3), 45-50.

10. Regenspurg, S.; Schild, D.; Schafer, T.; Huber, F.; Malmstrom, M. E., Appl. Geochem.

2009, 24 (9), 1617-1625.

11. Scott, T. B.; Allen, G. C.; Heard, P. J.; Randell, M. G., Geochim. Cosmochim. Ac. 2005,

69 (24), 5639-5646.

12. Scott, T. B.; Tort, O. R.; Allen, G. C., Geochim. Cosmochim. Ac. 2007, 71 (21), 5044-

5053.

Page 247: Relativistic Quantum Chemistry Applied to Actinides

225

13. Selli, E.; Eliet, V.; Spini, M. R.; Bidoglio, G., Environ. Sci. Technol. 2000, 34 (17), 3742-

3748.

14. Sherman, D. M.; Peacock, C. L.; Hubbard, C. G., Geochim. Cosmochim. Ac. 2008, 72

(2), 298-310.

15. Evans, C. J.; Nicholson, G. P.; Faith, D. A.; Kan, M. J., Green Chem. 2004, 6 (4), 196-

197.

16. Yusov, A. B.; Shilov, V. P., Russ. Chem. B+ 2000, 49 (2), 285-290.

17. Roberts, S. K.; Bourcier, W. L.; Shar, H. F., Radiochim. Acta 2000, 88, 539-543.

18. Helean, K. B.; Ushakov, S. V.; Brown, C. E.; Navrotsky, A.; Lian, J.; Ewing, R. C.;

Farmer, J. M.; Boatner, L. A., J. Solid State Chem. 2004, 177, 1858-1866.

19. Helean, K. B.; Navrotsky, A.; Lumpkin, G. R.; Colella, M.; Lian, J.; Ewing, R. C.;

Ebbinghaus, B.; Catlano, J. G., J. Nucl. Mater. 2003, 320, 231-244.

20. Fayek, M.; Ren, M.; Goodell, P.; Dobson, P.; Saucedo, A. L.; Kelts, A.; Utsunomiya, S.;

Ewing, R. C.; Riciputi, L. R.; Reyes, I., International High Level Radioactive Waste

Management Conference Proc. 2006.

21. Payne, T. E.; Davis, J. A.; Lumpkin, G. R.; Chisari, R.; Waite, T. D., Appl. Clay Sci.

2004, 26 (1-4), 151-162.

22. Perron, H.; Domain, C.; Roques, J.; Drot, R.; Simoni, E.; Catalette, H., Radiochim. Acta

2006, 94 (9-11), 601-607.

23. Perron, H.; Domain, C.; Roques, J.; Drot, R.; Simoni, E.; Catalette, H., Inorg. Chem.

2006, 45 (17), 6568-6570.

24. Perron, H.; Roques, J.; Domain, C.; Drot, R.; Simoni, E.; Catalette, H., Inorg. Chem.

2008, 47 (23), 10991-10997.

Page 248: Relativistic Quantum Chemistry Applied to Actinides

226

25. Hattori, T.; Saito, T.; Ishida, K.; Scheinost, A. C.; Tsuneda, T.; Nagasaki, S.; Tanaka, S.,

Geochim. Cosmochim. Ac. 2009, 73 (20), 5975-5988.

26. Kremleva, A.; Krüger, S.; Rösch, N., Radiochim. Acta 2010, 98 (9-11), 635-646.

27. Roques, J.; Veilly, E.; Simoni, E., Int. J. Mol. Sci. 2009, 10 (6), 2633-2661.

28. Veilly, E.; Roques, J.; Jodin-Caumon, M. C.; Humbert, B.; Drot, R.; Simoni, E., J. Chem.

Phys. 2008, 129 (24).

29. Usubharatana, P.; McMartin, D.; Veawab, A.; Tontiwachwuthikul, P., Ind. Eng. Chem.

Res. 2006, 45 (8), 2558-2568.

30. Bowker, M.; Bennett, R. A., J. Phys-Condes. Mat. 2009, 21 (47).

31. Deskins, N. A.; Rousseau, R.; Dupuis, M., J. Phys. Chem. C 2009, 113 (33), 14583-

14586.

32. Deskins, N. A.; Rousseau, R.; Dupuis, M., J. Phys. Chem. C 2010, 114 (13), 5891-5897.

33. Diebold, U., Surf. Sci. Rep. 2003, 48 (5-8), 53-229.

34. Finazzi, E.; Di Valentin, C.; Pacchioni, G.; Selloni, A., J. Chem. Phys. 2008, 129 (15).

35. Kresse, G.; Furthmüller, J., Comput. Mat. Sci 1996, 6, 15.

36. Kresse, G.; Furthmüller, J., Phys. Rev. B 1996, 54, 11169.

37. Kresse, G.; Hafner, J., Phys. Rev. B 1993, 47, 558.

38. Kresse, G.; Hafner, J., Phys. Rev. B 1994, 49, 14251.

39. Blöchl, P. E., Phys. Rev. B 1994, 50, 17953.

40. Kresse, G.; Joubert, D., Phys. Rev. B 1999, 59, 1758.

41. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1996, 77 (18), 3865-3868.

42. Perdew, J. P.; Burke, K.; Ernzerhof, M., Phys. Rev. Lett. 1997, 78 (7), 1396-1396.

43. Calzado, C. J.; Hernandez, N. C.; Sanz, J. F., Phys. Rev. B 2008, 77 (4).

Page 249: Relativistic Quantum Chemistry Applied to Actinides

227

44. Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P., Phys.

Rev. B 1998, 57, 1505

45. Morgan, B. J.; Watson, G. W., J. Phys. Chem. C 2010, 114 (5), 2321-2328.

46. Klamt, A., J. Phys. Chem. 1995, 99 (7), 2224-2235.

47. Klamt, A.; Schuurmann, G., J. Chem. Soc. Perk T. 2 1993, (5), 799-805.

48. Pye, C. C.; Ziegler, T., Theor. Chem. Acc. 1999, 101 (6), 396-408.

49. Faas, S.; Snijders, J. G.; van Lenthe, J. H.; van Lenthe, E.; Baerends, E. J., Chem. Phys.

Lett. 1995, 246 (6), 632-640.

50. van Lenthe, E., J. Comput. Chem. 1999, 20 (1), 51-62.

51. van Lenthe, E.; Baerends, E. J.; Snijders, J. G., J. Chem. Phys. 1993, 99 (6), 4597-4610.

52. ADF2010.01, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands,

http://www.scm.com 2010.

53. te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.; Fonseca Guerra, C.; Baerends,

E. J.; Snijders, J. G.; Ziegler, T., J. Comput. Chem. 2001, 22, 931-967.

54. Zhang, H. Z.; Banfield, J. F., J. Mater. Chem. 1998, 8 (9), 2073-2076.

55. Pan, Q.-J.; Odoh, S. O.; Asaduzzaman, A. M.; Schreckenbach, G., Chem-Eur. J.

accepted.

56. Purton, J.; Bullett, D. W.; Oliver, P. M.; Parker, S. C., Surf. Sci. 1995, 336 (1-2), 166-

180.

57. Ramamoorthy, M.; Vanderbilt, D.; Kingsmith, R. D., Phys. Rev. B 1994, 49 (23), 16721-

16727.

58. Reyes-Coronado, D.; Rodriguez-Gattorno, G.; Espinosa-Pesqueira, M. E.; Cab, C.; de

Coss, R.; Oskam, G., Nanotechnology 2008, 19 (14).

Page 250: Relativistic Quantum Chemistry Applied to Actinides

228

59. Jedidi, A.; Markovits, A.; Minot, C.; Bouzriba, S.; Abderraba, M., Langmuir 2010, 26

(21), 16232-16238.

60. Li, S. C.; Wang, J. G.; Jacobson, P.; Gong, X. Q.; Selloni, A.; Diebold, U., J. Am. Chem.

Soc. 2009, 131 (3), 980-984.

61. Morgan, B. J.; Watson, G. W., J. Phys. Chem. C 2009, 113 (17), 7322-7328.

62. Bonapasta, A. A.; Filippone, F.; Mattioli, G.; Alippi, P., Catal. Today 2009, 144 (1-2),

177-182.

63. Henderson, M. A.; Epling, W. S.; Peden, C. H. F.; Perkins, C. L., J. Phys. Chem. B 2003,

107 (2), 534-545.

64. Wendt, S.; Matthiesen, J.; Schaub, R.; Vestergaard, E. K.; Laegsgaard, E.; Besenbacher,

F.; Hammer, B., Phys. Rev. Lett. 2006, 96 (6).

65. Den Auwer, C.; Drot, R.; Simoni, E.; Conradson, S. D.; Gailhanou, M.; de Leon, J. M.,

New J. Chem. 2003, 27 (3), 648-655.

66. Bondarchuk, O.; Kim, Y. K.; White, J. M.; Kim, J.; Kay, B. D.; Dohnalek, Z., J. Phys.

Chem. C 2007, 111 (29), 11059-11067.

67. Comarmond, M. J.; Payne, T. E.; Harrison, J. J.; Thiruvoth, S.; Wong, H. K.; Aughterson,

R. D.; Lumpkin, G. R.; Muller, K.; Foerstendorf, H., Environ. Sci. Technol. 2011, 45 (13), 5536-

5542.

68. Dorado, B.; Amadon, B.; Freyss, M.; Bertolus, M., Phys. Rev. B 2009, 79 (23).

69. Kotani, A.; Yamazaki, T., Prog. Theor. Phys. Suppl. 1992, (108), 117-131.

70. Shapovalov, V.; Stefanovich, E. V.; Truong, T. N., Surf. Sci. 2002, 498 (1-2), L103-

L108.

71. Bakac, A.; Espenson, J. H., Inorg. Chem. 1995, 34 (7), 1730-1735.

Page 251: Relativistic Quantum Chemistry Applied to Actinides

229

72. Sarakha, M.; Bolte, M.; Burrows, H. D., J. Phys. Chem. A 2000, 104 (14), 3142-3149.

73. Nagaishi, R.; Katsumura, Y.; Ishigure, K.; Aoyagi, H.; Yoshida, Z.; Kimura, T.; Kato, Y.,

J. Photoch. Photobio. A. 2002, 146 (3), 157-161.

74. Tsushima, S., Inorg. Chem. 2009, 48 (11), 4856-4862.

75. Moskaleva, L. V.; Krüger, S.; Sporl, A.; Rösch, N., Inorg. Chem. 2004, 43 (13), 4080-

4090.

76. Vallet, V.; Maron, L.; Schimmelpfennig, B.; Leininger, T.; Teichteil, C.; Gropen, O.;

Grenthe, I.; Wahlgren, U., J. Phys. Chem. A 1999, 103 (46), 9285-9289.

77. Vallet, V.; Schimmelpfennig, B.; Maron, L.; Teichteil, C.; Leininger, T.; Gropen, O.;

Grenthe, I.; Wahlgren, U., Chem. Phys. 1999, 244 (2-3), 185-193.

78. Formosinho, S. J., J. Chem. Soc. Faraday T. II 1976, 72, 1313-1331.

Page 252: Relativistic Quantum Chemistry Applied to Actinides

230

Preface to Chapter 7

This chapter is based on a manuscript published in the journal “Inorganic Chemistry”. The full

citation of the paper is as follows:

Samuel O. Odoh, Sean M. Walker, Markus Meier, Jörg Stetefeld and Georg

Schreckenbach, “QM and QM/MM Study of Uranyl Fluorides in the Gas Phase, Aqueous

Phase and in the Hydrophobic Cavities of Tetrabrachion”, Inorganic Chem. 2010, 50,

3141-3152.

The existence, orientation and electronic properties of an uranyl fluoride complex (UO2F53-

) in

the hydrophobic cavities of tetrabrachion protein unit were examined using density functional

theory and molecular mechanics in a QM/MM formalism.

All the calculations in the published manuscript and compiled in this chapter were carried out by

Samuel O. Odoh. The manuscript was prepared together with the other authors.

Copyright permissions have been obtained from the American Chemical Society and the other

authors. Sean M. Walker was an undergraduate research fellow in Professor Schreckenbach‟s

group. Dr Markus Meier and Dr Jörg Stetefeld are affiliated with the Department of Chemistry at

University of Manitoba.

Page 253: Relativistic Quantum Chemistry Applied to Actinides

231

Chapter 7: QM and QM/MM Study of Uranyl Fluorides in the Gas

Phase, Aqueous Phase and in the Hydrophobic Cavities of

Tetrabrachion

Abstract

The structural properties and electronic structures of pentacoordinated uranyl complexes

belonging to the [UO2Fn(H2O)5-n]2-n

series have been studied in the gas and aqueous phases using

density functionals with relativistic pseudopotentials and all electron basis sets in the gas phase

calculations in combination with the COSMO solvation model in the aqueous phase. In addition,

the conformational orientation, structural and electronic properties of [UO2F5]3-

in the

hydrophobic cavities of the right handed coiled coil (RHCC) protein of tetrabrachion have been

determined using the hybrid QM/MM method. Although, there is good agreement between the

available experimental geometrical parameters and the values obtained in the aqueous phase

using pseudopotentials or all electron basis sets, the variation of the uranyl U=O bond with the

number of fluoride ligands is only truly captured after the inclusion of five water molecules in

the second coordination sphere around the molecules. The docking procedure used in this work

shows that there are only two possible orientations of the uranyl group of [UO2F5]3-

embedded in

the hydrophobic cavities of the RHCC protein. The two orientations are exclusively along the

axes perpendicular to the protein axial channel with no possible orientation of the uranyl group

along the axial channel due to both steric effects and interaction with the alkyl chain of the

isoleucine residues pointing into the axial channel. In addition, the embedded complex is always

positioned nearer the isoleucine residues at the N-terminal ends of the hydrophobic cavities.

Page 254: Relativistic Quantum Chemistry Applied to Actinides

232

Energy analysis however reveals that both conformations can only be observed in cavity two, the

largest hydrophobic cavity. The structural and electronic properties of the ligand embedded in

this cavity are very similar to those of the gas-phase structure. Comparable study of [Pt(CN)6]2-

and the anticancer drug, cisplatin, [PtCl2(NH3)2], in cavity two, revealed the existence of just two

orientations for the former, similar to the uranyl complex and multiple orientations for the latter.

Introduction

The tetrabrachion complex of archaebacterium Staphylothermus marinus has been shown

to contain a right-handed coiled coil (RHCC) protein.1-5

The thermodynamically stable RHCC

contains four parallel α-helical chains oriented in a right-handed fashion with four hydrophobic

cavities aligned along the axis of the protein, Figure 7.1. In the native crystal structure, the

largest hydrophobic cavity of tetrabrachion contains nine water molecules aggregated into a

cluster while the other cavities contain five, one and two water molecules respectively according

to their sizes. The aggregation of the water molecules in the cavities of tetrabrachion is due to the

exclusively aliphatic and hydrophobic lining of the cavity walls. Yin et al have shown that

metastable water complexes held together by hydrogen bonds exist in the largest cavity of RHCC

both at room (298 K) and high (365 K) temperatures.6-7

They also demonstrated the existence of

significant entropic contributions to the thermodynamics of the filling of the largest hydrophobic

cavity by multiple water molecules.

A recent review on the nature and structure of coiled-coil proteins as well as their

potential use for therapeutic purposes has been published by McFarlane et al.8 The presence of

large cavities in such proteins and the ability of these cavities to bind or hold cargo molecules

make them ideal drug delivery vehicles. Modification of the terminal amino acids of coiled coils

with specific labeling groups can be used for specific drug targeting thus reducing the overall

Page 255: Relativistic Quantum Chemistry Applied to Actinides

233

cytotoxicity of a therapeutic molecule. Perhaps the most important potential use of coiled-coil

proteins is for delivery of cancer drugs to tumor cells. In fact, the ability of the RHCC

Figure 7.1: Top: The Right Handed Coiled Coil Protein of Tetrabrachion. Cavities One-Four

from the N-Terminus (Left) to the C-Terminus (Right) are Shown. Bottom: Two Monomer

Chains of the RHCC Tetramer. Isoleucine and Leucine Side Chains are Found at the N-Terminal

and C-Terminal Ends Respectively of Cavities Two and Three.

Leucine Isoleucine

Page 256: Relativistic Quantum Chemistry Applied to Actinides

234

of tetrabrachion to incorporate and transfer cisplatin, [PtCl2(NH3)2], into mammalian cells has

been studied by Eriksson et al.2 They found that RHCC stably incorporates cisplatin at room

temperature and that the RHCC-cisplatin (RHCC-C) complex rather efficiently binds to cells.

The RHCC-C complex was found to be equally or sometimes more effective against cancer cells

as the pure cisplatin drug. Their work raises the possibility of the RHCC protein being used as a

carrier for cisplatin in therapeutic usage. However, an obstacle to the long-term goal of using the

RHCC as a drug delivery vehicle is the issue of diseased cell targeting. Theoretical studies of

cisplatin in the cavities of the RHCC could potentially be used in designing modifications to

either the cargo molecule or the aliphatic chains lining the cavities.

Moreover, cancer drugs are not the only kind of molecular systems that can be embedded

in the hydrophobic cavities of RHCC. Indeed several heavy metals and their compounds were

incorporated into the cavities of the tetrabrachion RHCC during X-ray crystallographic studies of

its structure.5 It would thus appear that the hydrophobic cavities are filled with water clusters in

the native structure while the occupying water clusters are displaced by any compact molecular

or ionic system present in solution. The displacement of water clusters embedded in the

hydrophobic cavities by a single ionic complex will be favored by an increase in total entropy

which will dominate the positive-leaning enthalpy of transferring ionic or hydrophilic species

from the polar aqueous solvent into a cavity lined exclusively with aliphatic side chains.6-7

Actinide complexes like uranyl fluoride used in the determination of the phase

information of the RHCC are also incorporated into the hydrophobic cavities.4-5, 9-10

Uranyl

fluorides have been extensively studied theoretically and experimentally in both the gaseous and

aqueous phases.11-26

The hydrophobic cavities of the RHCC of tetrabrachion represent a rather

unique and „different‟ environment in which to study the structure and bonding of uranyl

Page 257: Relativistic Quantum Chemistry Applied to Actinides

235

fluorides. The extent to which the wavefunction of a cargo molecule is perturbed by the protein

environment can be determined by comparing the geometrical and electronic structures of uranyl

fluorides in the hydrophobic cavities of RHCC, in the gas phase and in solution. In addition, as

the cavities of the RHCC are non-polar, it will be interesting to see if there is any „local-order‟

(or favored orientations) to the alignment of polar molecules like uranyl fluoride.

Full quantum mechanical (QM) calculations on the protein-actinide species complex are

currently very computationally expensive not in the least because of the large number of degrees

of freedom in the protein.27

Indeed, the large number of loose degrees of freedom in biological

macromolecules makes the concept of a „global or local‟ structural minimum less important than

in stiff molecules and hence the need for selection of a probabilistic ensemble corresponding to

all possible configurations in which the macromolecule could exist at a certain temperature.27

The quantum mechanical/molecular mechanics (QM/MM) method is a hybrid method in which

the active or interesting site is treated with computationally more demanding QM methods while

the remainder of the system (remainder of the macromolecule and/-or environmental water) is

treated classically using molecular mechanics (MM).28-42

This method has been used extensively

in literature to study large systems and represents a balance between the accuracy of full

quantum mechanical treatment and the computational efficiency of a full classical treatment

using molecular mechanics.43-46

There have been very few theoretical studies of actinide complexes using the QM/MM

method. Infante et al studied the nature of the water solvation shells around both the tetrafluoro-

and tetrahydroxo- complexes of the uranyl dication.15-16, 47-48

The solvent water molecules were

treated with MM while the uranyl complexes were treated with density functional theory (DFT)

with relativistic effects included using the zeroth order regular approximation, ZORA.49-51

The

Page 258: Relativistic Quantum Chemistry Applied to Actinides

236

interaction between the solvent water MM region and the wavefunction of the actinide complex

QM region was restricted to mechanical coupling. Comparison of the results of QM/MM

calculations with those obtained with full QM calculations indicated qualitative and some

quantitative agreements in the computed geometric and electronic properties.

Here we present QM/MM simulation studies on uranyl pentafluoride, [UO2F5]3-

incorporated in the hydrophobic cavities of the RHCC using DFT with relativistic effective core

potentials (RECPs) to include relativistic effects on the QM region while classically representing

the RHCC using the popular AMBER95 force field.52

Electrostatic coupling of the MM charges

to the wavefunction of the QM region and other non-bonded interactions like Van der Waals and

electrostatic interactions are included in the calculations. The structural and electronic properties

of [UO2F5]3-

in the hydrophobic cavities are compared to those in both gaseous and aqueous

phases. The presence of favored configurations for the embedded actinide complex with respect

to the axial channel of the protein is reliably proven using the QM/MM method. This is in

agreement with preliminary experimental evidence for the existence of two orientations for

uranyl fluoride in the largest hydrophobic cavity of the RHCC.53

Finally, the incorporation of the

anticancer drug, cisplatin, and [Pt(CN)6]2-

into the largest hydrophobic cavity is examined using

the same methodology used for uranyl fluoride. It should be fully noted that current theoretical

calculations of cisplatin in the hydrophobic cavities were stimulated by the possibility of

therapeutic use while similar work on [UO2F5]3-

in these cavities was motivated by structural and

electronic considerations (the presence of local order in the arrangement of this molecule in the

cavities and the effect of the cavity walls on the electronic structure of the actinide complex).

The common theme connecting the two subjects (cisplatin and uranyl complexes) is provided in

the common protein environment with its unique cavity structure and the resulting application of

Page 259: Relativistic Quantum Chemistry Applied to Actinides

237

a common methodology. There is undoubtedly little therapeutic potential for this actinide

complex.

The remainder of this report is organized as follows. The computational calculations

carried out are first described followed by a discussion of the geometric and electronic structure

of uranyl fluorides in the gas and aqueous phases. The chemistry of uranyl fluorides in the

hydrophobic cavities of tetrabrachion is then described using [UO2F5]3-

as a representative

complex. Finally, we compare the incorporation of [UO2F5]3-

in the hydrophobic cavities to that

of other ligands such as cisplatin and [Pt(CN)6]2-

.

Computational Details

The molecular geometries of all members of the [UO2Fn(H2O)5-n]2-n

series and cisplatin

were optimized in the gaseous and aqueous phases using the B3LYP54-57

and BP8657-58

functionals. The uranium and platinum atoms were described with the Stuttgart small-core (60

core electrons represented by a pseudopotential) RECP and associated valence basis sets while

all other atoms were described with the 6-311++G** basis set.59-61

All g-functions in the valence

basis associated with the Stuttgart pseudopotentials were removed. In addition, a set of diffuse f-

type basis functions (α = 0.005) was added to allow for an accurate description of the lowest

unoccupied molecular orbitals of the uranyl complexes. All the RECP calculations were carried

out in the NWChem 5.1.1 package.62-63

The aqueous phase calculations employed the conductor-

like screening solvation model (COSMO)64-65

. The atomic radii used in forming solvation

cavities around the molecules in these calculations are 2.18, 1.72, 1.72 and 1.30 for the uranium,

fluorine, oxygen and hydrogen atoms, respectively.

Page 260: Relativistic Quantum Chemistry Applied to Actinides

238

Also the ZORA relativistic approach49-51

with triple-ζ polarized (TZP) all electron basis

sets and the BP86 functional was used in optimizing the geometries of all the molecules. No core

atomic orbitals were frozen. These ZORA-DFT calculations were carried out using the

Amsterdam Density Functional (ADF 2009) package with an integration parameter of 6.0.66-68

Multipole derived atomic charges, Mayer bond orders and Mulliken atomic charges were

obtained from these all-electron calculations.69

In ADF,70

the aqueous phase calculations were

carried out using the COSMO solvation model64-65

and identical atomic radii to those used in the

RECP calculations were employed. Modern approaches in theoretical calculations of actinide

chemistry in the gaseous and aqueous phases have recently been reviewed.71

The starting configurations for the QM/MM simulations of [UO2F5]3-

, [Pt(CN)6]2-

and

cisplatin encapsulated in the hydrophobic cavities of the RHCC were generated using

AutoDock.72

The water clusters, (H2O)n (where n=9, 5, 1 and 2 for the largest, second-largest,

third-largest and smallest cavities, respectively) in the cavities were evacuated prior to docking.5

Ligand-optimized geometries and multipole-derived atomic charges obtained from the BP86-

ZORA-TZP calculations were used in the docking. The van der Waals parameters used for the

uranium atom were taken from the work of Guilbaud and Wipff.73

The parameters used for

platinum atoms when cisplatin was docked in the cavity were obtained from the work of Spiegel

et al.74

The hydrophobic cavities or binding sites for the ligands were determined from the amino

acid sequence of the X-ray structure3 and confirmed using various grids in AutoDock. The

simulated annealing algorithm was used to generate structures, which were then clustered and

ranked by their energies. Several initial temperatures for the annealing were tried in addition to

using 25, 50 and 75 cooling runs.

Page 261: Relativistic Quantum Chemistry Applied to Actinides

239

The atoms of the RHCC protein and crystal waters were represented with the AMBER95

force field in the QM/MM calculations.52

The QM/MM calculations were performed using the

NWChem 5.1.1 and 6.0 codes.62-63

There was no necessity for link atoms as there are no formal

bonds between the embedded molecules and the cavity walls of the protein. DFT calculations

using the BP86 functional as well as the Stuttgart small-core RECP for the uranium and platinum

atoms and the 6-311++G** basis for all other atoms were performed on the embedded

molecules.61

The cutoff for all non-bonded interactions was set at 15 Å. The ligands considered

in this work are [UO2F5]3-

and cisplatin embedded in the protein hydrophobic cavities.

Electrostatic coupling (polarization of the wave function of the QM region by the charges of

neighboring MM atoms), Coulombic electrostatic and van der Waal (vdW) interactions of the

protein and ligand atoms were included in the calculations. The geometry optimization procedure

included sequential optimization of both the embedded ligands (QM) and the RHCC protein

(MM) till energy convergence (5.0×10-5

Hartree) was attained. The vibrational frequencies of the

embedded ligands were also determined after the geometry optimization by numerical

differentiation. All the MM atoms were held frozen during the finite-difference vibrational

calculations.

Results and Discussion

[UO2Fn(H2O)5-n]2-n

Complexes: Gaseous and Aqueous Phases. The structural parameters and

uranyl vibrational stretching frequencies of the [UO2Fn(H2O)5-n]2-n

compounds, Figure 7.2,

obtained using DFT with the small-core RECP and the all-electron ZORA approaches are given

in Table 7.1. Methodologically, the U=O bond lengths computed using the B3LYP hybrid

functional are generally shorter than those obtained with the BP86 functional. This agrees with

literature experience, Table 7.1.21, 75-77

Also, given comparable basis set size and the same

Page 262: Relativistic Quantum Chemistry Applied to Actinides

240

density functional, nearly identical bond lengths and vibrational frequencies are obtained from

the RECP and ZORA approaches, Table 7.1. This agrees with recent estimates that the bond

lengths obtained with small-core RECPs agree with those obtained using a four component

relativistic approach while the calculated vibrational wavenumbers are lower than the all-

electron basis set results.21, 75

Figure 7.2: Aqueous Phase Structures of the [UO2Fn(H2O)5-n]2-n

Complexes Optimized at the

BP86/TZP/ZORA/COSMO Level. A) [UO2F5]3-

, B) [UO2(H2O)F4]2-

Structure 1, C)

[UO2(H2O)F4]2-

Structure 2 D) [UO2(H2O)2F3]1-

Structure 1, E) [UO2(H2O)2F3]1-

Structure 2 F)

[UO2(H2O)3F2]0 Structure 1, G) [UO2(H2O)3F2]

0 Structure 2 H) [UO2(H2O)4F]

1+ and I)

[UO2(H2O)5]2+

.

B A C

D E F

H G

G

I

Page 263: Relativistic Quantum Chemistry Applied to Actinides

241

Table 7.1a: Calculated and experimental structural parameters of the [UO2Fn(H2O)5-n]2-n

complexes obtained at the RECP/BP86

level. The ADF/ZORA/BP86/TZP results values are given in parenthesis. The experimental data are from Reference 24

.

[UO2(H2O)5]2+

[UO2(H2O)5]2+

.5H2O Expt.b

Gas Aqueous Gas Aqueous

B3LYP BP86 a B3LYP BP86

a B3LYP BP86 B3LYP BP86

U=O 1.748 1.773 (1.772) 1.758 1.783(1.787) 1.759 1.789(1.789) 1.768 1.800(1.797) 1.76

U-OH2 2.499 2.481(2.494) 2.469 2.458(2.450) 2.466 2.464(2.468) 2.428 2.431(2.439) 2.41

νsymm 927 864 (891) 898 840 (859) 899 843(867) 878 870

νasymm 1015 958 (883) 960 911(922) 991 922(940) 939 965

[UO2(H2O)4F]1+

Gas Aqueous

B3LYP BP86 B3LYP BP86

U=O 1.771 1.795 (1.798) 1.776 1.802(1.807)

U-OH2 2.548 2.539(2.557) 2.515 2.510(2.510)

U-F 2.107 2.104(2.091) 2.146 2.134(2.128)

νsymm 881 827(846) 863 808(825)

Page 264: Relativistic Quantum Chemistry Applied to Actinides

242

νasymm 962 910(927) 917 869(877)

[UO2(H2O)3F2]0 Structure 1 [UO2(H2O)3F2]

0 Structure 2

Gas Aqueous Gas c Aqueous

c

B3LYP BP86 B3LYP BP86 B3LYP

[3.4]

BP86 [2.5] B3LYP

[1.6]

BP86 [0.02]

U=O 1.785 1.814 (1.813) 1.792 1.820 (1.826) 1.783 1.809 (1.809) 1.792 1.820 (1.825)

U-OH2 2.666 2.657 (2.663) 2.652 2.579 (2.577) 2.594 2.594 (2.605) 2.559 2.562 (2.568)

U-F 2.157 2.150 (2.154) 2.167 2.165 (2.160) 2.180 2.170 (2.210) 2.183 2.170 (2.164)

νsymm 851 798 (824) 831 779 (793) 856 806 (826) 831 779 (793)

νasymm 928 876 (898) 881 831 (835) 935 886 (905) 879 832 (838)

Page 265: Relativistic Quantum Chemistry Applied to Actinides

243

Table 7.1b: Calculated and experimental structural parameters of the [UO2Fn(H2O)5-n]2-n

complexes obtained at the RECP/BP86

level. The ADF/ZORA/BP86/TZP results values are given in parenthesis. The experimental data are from Reference 24

.

[UO2(H2O)2F3]1-

Structure 1 [UO2(H2O)2F3]1-

Structure 2 Expt.

Gas Aqueous Gas c Aqueous

c

B3LYP BP86 B3LYP BP86 B3LYP,8.6 BP86,9.4 B3LYP,6.6 BP86,6.9

U=O 1.805 1.841 (1.842) 1.803 1.834 (1.838) 1.799 1.828 (1.828) 1.807 1.832 (1.842) 1.80

U-OH2 2.620,

3.715

2.594,3.614

(2.605,3.658)

2.566,

4.172

2.588,4.033

(2.507,3.888)

2.776 2.804 (2.790) 2.610 2.625 (2.625) 2.47

U-F 2.192 2.183 (2.183) 2.215 2.196 (2.203) 2.221 2.211 (2.219) 2.221 2.226 (2.203) 2.25

νsymm 812 777 (783) 807 761 (777) 831 814 (822) 805 761 (767)

νasymm 884 832 (869) 854 809 (814) 901 850 (853) 848 809 (804)

[UO2(H2O)F4]2-

Structure 1 [UO2(H2O)F4]2-

Structure 2

Gas Aqueous Gas c Aqueous

c

B3LYP BP86 B3LYP BP86 B3LYP,34.7 BP86,35.1 B3LYP,11. BP86 ,11.9

U=O 1.820 1.850 1.819 1.850 1.824 1.856 1.824 1.857 (1.860) 1.80

U-OH2 4.022 4.003 3.958 3.917 2.710 2.701 2.672 2.685 (2.715) 2.48

Page 266: Relativistic Quantum Chemistry Applied to Actinides

244

U-F 2.250 2.240 2.230 2.220 2.260 2.270 2.250 2.244 (2.248) 2.26

νsymm 789 741 (757) 786 737 (755) 776 727 778 726 (740)

νasymm 861 815 (830) 825 778 (784) 845 794 814 765(771)

[UO2F5]3-

[UO2F5]3-

.5H2O

Gas Aqueous Gas Aqueous

B3LYP BP86 B3LYP BP86 B3LYP BP86 B3LYP BP86

U=O 1.842 1.878 (1.876) 1.835 1.871 (1.874) 1.804 1.832(1.831) 1.811 1.837(1.840) 1.80

U-F 2.343 2.337 (2.338) 2.300 2.294 (2.281) 2.365 2.361(2.370) 2.328 2.330(2.326) 2.26

νsymm 740 688 (694) 757 705(716) 803 754(775) 799 751(764) 784

νasymm 806 752 (768) 790 739(745) 865 825(843) 836 788(808) 850

Page 267: Relativistic Quantum Chemistry Applied to Actinides

245

There is a gradual increase in the calculated U=O and U-F bond lengths as the number of

fluoride ligands is increased down the [UO2Fn(H2O)5-n]2-n

series, Table 7.1. This is especially

true for conformers in which the five ligands are in the first coordination sphere about the

equatorial plane. The calculated Mayer bond orders of the U=O bonds in the complexes are

presented in Table 7.2. Concomitant with the increasing U=O bond lengths, there is a decrease in

the calculated U=O bond order as the number of fluoride ligands in the complexes is increased.

This lengthening of the U=O bond is accompanied by a decrease in the calculated Mulliken

charges on both the uranium atom and the uranyl group, Figure 7.3 and Table 7.2. Based on the

calculated Mayer bond orders and atomic charges in the molecules, an additive-ionic Lewis base

effect of multiple fluoride ligands could be used as an explanation for the increasing U-F bonds13

even as such an approach can be used in explaining the increase in the U=O and U-OH2 bond

lengths.

The sequential-average ligand binding energies of the uranyl complexes provide an

alternative way of examining the effect of greater number of fluoride ligands. The ligand binding

energy, ΔEbinding, to the uranyl moiety of a [UO2Fn(H2O)5-n]2-n

complex is given by Ecomplex –

Euranyl – nEfluoride ion – (5-n)Ewater. The difference between the ligand binding energies of

successive members of the [UO2Fn(H2O)5-n]2-n

series represents the average energy required for

the sequential replacement of an aquo ligand by a fluoride ion. The average ligand binding

energies calculated for the [UO2Fn(H2O)5-n]2-n

complexes in solution increases as the number of

fluoride ligands increases, Table 7.2. This is to be expected given the replacement of a neutral

aquo ligand coordinated to the uranyl cation by an anionic fluoride ligand. However the

calculated energies for introduction of a subsequent fluoride anion reduces from -43.8 kcal/mol

Page 268: Relativistic Quantum Chemistry Applied to Actinides

246

in the case of [UO2(H2O)5]2+

to 3.70 kcal/mol in the case of [UO2(H2O)F4]1+

. Further

examination Table 7.2 shows that the calculated successive ligand binding energies overestimate

Figure 7.3: Left: Variation of the Mulliken Charges on the Uranium Atom (Black) and Uranyl

Moiety (Red) and the Fluoro-2p Contribution to the HOMO-1 Orbital (Blue) with Increasing

Number of Fluoride Ligands. Right: Variation of the U=O Bond Lengths (Black) and Bond

Orders (Red) with Increasing Number of Fluoride Ligands.

the experimental values significantly, likely a result of the poor performance of the solvation

model for calculating ligand binding and activation energies.12, 24, 78

The use of extended second

and third aquo-coordination spheres in addition to the implicit PCM model around the uranyl

complexes would lead to improvements in the calculated ligand binding energies.79

Electronically, the first few virtual molecular orbitals in all the uranyl compounds are all

of uranium 5f character, Figures 7.4a and 7.4b. This is generally expected for 5f0 U(VI)

complexes. The highest occupied orbitals (HOMO) are all uranyl based and are almost identical

Page 269: Relativistic Quantum Chemistry Applied to Actinides

247

Table 7.2: Aqueous phase calculated ligand binding energies in kcal/mol, Mayer bond orders for

the U=O Bond and atomic Mulliken charges on uranium atoms in the UO2Fn(H2O)5-n]2-n

complexes obtained at the ADF/ZORA/TZP/BP86/COSMO level.

n Calculated

Ligand B. E.b

Experimental

Successive

Ligand B. E.86

Mulliken Charges on

Uranium Atom

U=O Mayer Bond

Orders

Gas Aqueous Gas Aqueous

0 -95.8 2.364 2.482 2.150 2.065

1 -139.6 (-43.8) -7.04 2.304 2.418 2.105 2.032

2 -176.2 (-36.6) -5.00 2.286 2.330 2.060 1.989

2a -176.0 (-36.4) 2.321 2.319 2.060 1.982

3 -205.2 (-29.0) -2.82 2.215 2.257 1.987 1.935

3a -210.5 (-34.3) 2.260 2.275 2.000 1.960

4 -234.3 (-23.8) -1.28 2.235 2.241 1.944 1.920

4a -222.9 (-17.8) 2.200 2.154 1.907 1.889

5 -230.6 (+3.70) 1.00 2.094 2.095 1.854 1.844

a These are structures 2 for the species with n=2, 3 and 4 respectively, Figure 2 and text.

in all optimized structures of the [UO2Fn(H2O)5-n]2-n

compounds. They are of σ-character with 5f

atomic contribution from the uranium atom and 2p-contributions from the oxygen atoms of the

uranyl group, Figures 7.4c and 7.4d. Contributions from the 6p-atomic orbitals of uranium are

also

Page 270: Relativistic Quantum Chemistry Applied to Actinides

248

Figure 7.4: Selected Frontier Molecular Orbitals of [UO2Fn(H2O)5-n]2-n

Member Structures

Optimized at the ADF/ZORA/BP86/TZP/COSMO Level: A) LUMO of [UO2(H2O)2F3]1-

Structure 2 B) LUMO of [UO2(H2O)4F]1+

C) HOMO of [UO2(H2O)5]2+

D) HOMO of [UO2F5]3-

,

E) HOMO-1 of [UO2(H2O)F4]2-

and F) HOMO-1 of [UO2(H2O)3F2]0 Structure 2.

found in the HOMO σ-orbitals for these complexes. However, there are sometimes very minor

contributions from the ligand 2p orbitals in a π-character to the HOMO in these complexes. The

HOMO-1 to HOMO-4 orbitals in all the complexes are generally π-type orbitals formed from the

2p-type atomic orbitals of the equatorial ligands as well as in most cases 2p-orbital contributions

from the uranyl oxo atoms in a π-bonding scheme. Contributions from the uranium 6p and 6d-

orbitals to the HOMO to HOMO-4 orbitals are minor and amount to not more than 1.3% and

2.5% respectively. The increase in the number of fluoride ligands down the series can be

observed in the evolution of the fluorine 2p contributions to the HOMO-1, Figure 7.3. There is

A B C

D E F

Page 271: Relativistic Quantum Chemistry Applied to Actinides

249

scant evidence for any form of π-competition between the uranyl oxo-atoms and the equatorial

ligands. A more ionic Lewis base effect appears to be a more plausible explanation for the

increase in U=O and U-F bond lengths with increasing number of fluoride ligands in the

equatorial plane. In general, the HOMO-1 orbitals are similar in all the complexes except in

[UO2F5]3-

and [UO2(H2O)5]2+

for which there are no contributions to their HOMO-1 orbitals from

the uranyl oxo atoms, Figure 7.4.

Structurally, the [UO2F5]3-

complex belongs to the D5H symmetry group in both the gas

and aqueous phases (some slight loss in symmetry) in agreement with experimental solution24

and solid state26

observations. Optimization in solution results in a slight contraction of the U-F

bonds by approximately 0.04 Å while having little effect on the U=O bonds. The calculated

symmetric and asymmetric stretching vibrational frequencies of the uranyl group of [UO2F5]3-

deviate significantly from the experimental values.80-81

The discrepancy is most likely due to the

insufficient description of the aqueous environment around the ionic complex by the implicit

COSMO solvation model used. Indeed, better agreement was obtained after inclusion of five

water molecules in the second coordination sphere around the uranyl group, Table 7.1. The

presence of a second coordination sphere with five water molecules results in the shortening of

the U=O bond by 0.02-0.04 Å while causing an increase in the U-F bonds.

The U-OH2 bonds of the optimized C1 structure of [UO2(H2O)5]2+

are generally of two

types in aqueous solution: 4 water ligands arranged orthogonally to the equatorial plane and the

last aquo ligand almost parallel to the equatorial plane. This orientation is only slightly more

stable than the symmetrical D5 and D5H structures. Electronically, the introduction of D5 and D5H

high symmetries alters the orbital ordering in the complex. The major effect of imposing the D5

and D5H symmetries is a stabilization of the σ-type HOMO in the C1 structure relative to the π-

Page 272: Relativistic Quantum Chemistry Applied to Actinides

250

type HOMO-1. The calculated U=O and U-OH2 bond lengths in the minimum structure are in

good agreement with both experimental results and previous theoretical calculations.76, 82

Although the O=U=O bond angles in this molecule are slightly bent by 0.0-4.0º, it should be

noted that Perron et al have mentioned that the XANES data of [UO2(H2O)5]2+

could actually be

explained with O=U=O bond angles not less than 160 degrees.83

Similar to [UO2F5]3-

, Table 7.1,

the addition of five water molecules in the second coordination sphere results in better agreement

between the calculated and experimental uranyl stretching vibrational frequencies.

The addition of a second coordination sphere containing five water molecules to the

structures of [UO2(H2O)5]2+

and [UO2F5]3-

not only results in greater agreement between the

experimental and calculated vibrational frequencies, it allows the experimental range of the U=O

bond lengths in the [UO2Fn(H2O)5-n]2-n

complexes to be accurately captured by the theoretical

calculations, Table 7.1. The difference in the experimental U=O bond lengths of [UO2(H2O)5]2+

and [UO2F5]3-

is 0.04 Å in contrast to the 0.08-0.09 Å obtained using the COSMO solvation

model but in full agreement with the 0.037-0.043 Å using a combination of the COSMO model

with five water molecules in the second coordination sphere. This is particularly important given

that recent experimental work has revealed little evidence for variation in the U=O bond lengths

among the species in the [UO2Fn(H2O)5-n]2-n

series24

.

The structure of the [UO2(H2O)4F]1+

complex, Figure 7.2h, is a simple arrangement of the

oxygen atoms of the four water ligands and the fluoride ion in the equatorial plane. The uranyl

O=U=O angle in this complex ranges from 168.8º to 174.4º in the gaseous and aqueous phases.

Comparison to [UO2(H2O)5]2+

indicates a slight increase in the U=O (0.02 Å) and U-OH2 (0.045

Å) bond lengths accompanied by a subsequent decrease in the vibrational wavenumbers of the

U=O bond stretch. This increase in the U=O bond lengths (and also in the U-F bond lengths

Page 273: Relativistic Quantum Chemistry Applied to Actinides

251

down the series) is directly related to the decrease in the calculated charges on the uranium atom

and the uranyl group, Figure 7.3. The electron-donating fluoride ligand leads to the decrease in

charge on the actinide center resulting in weaker U-F bonds as more fluoride ligands are added

with an added effect of some electrostatic repulsion with the uranyl oxo atoms and equatorial

water ligands.13

There are many possible arrangements of the fluoro- and aquo-ligands in the complexes

intermediate between [UO2(H2O)4F]1+

and [UO2(H2O)F4]2-

. For [UO2(H2O)3F2]0, only the

structures in which all the aquo- and fluoro-ligands lie in the equatorial region were considered.

Structure 1 (or cis-structure with neighboring fluoro-ligands), Figure 7.2f, was calculated to be

approximately 3.4 and 1.6 kcal/mol less stable than structure 2 (or trans-structure with non-

neighboring fluoro-ligands), Figure 7.2g, at the B3LYP/6-311++G** level in the gas and

aqueous phases respectively. The energy difference between these structural isomers is however

0.02 and 0.12 kcal/mol at the RECP-BP86 and ZORA-BP86 levels respectively in aqueous

solution. The small magnitude of this energy difference might suggest co-existence or facile

inter-convertibility. In addition, the calculated geometrical parameters and vibrational

frequencies for both structures are identical especially in the aqueous phase. The only exception

to this is a U-OH2 bond in structure 1 which was calculated to be about 0.22 Å longer than all the

U-OH2 bonds in structure 2.

The structure corresponding to the alignment of three neighboring fluoro-ligands as well

as one aquo ligand in the equatorial plane represents the most stable conformation of the

[UO2(H2O)2F3]1-

complex, Structure 1, Figure 7.2d. The other aquo ligand is at a long-distance of

3.57-4.17 Å from the uranium atom. This long distance representing an aquo-ligand outside the

first coordination sphere was however not observed experimentally.24

Structure 1 was calculated

Page 274: Relativistic Quantum Chemistry Applied to Actinides

252

to be approximately 6.6-9.4 kcal/mol more stable than the structure corresponding to two non-

neighboring aquo ligands with the fluoro-ligands aligned in a triangular fashion at the equatorial

plane, Structure 2, Figure 7.2e. The calculated U=O and U-F bond lengths for Structure 2 are in

good agreement with experimental data in contrast to the U-OH2 bonds which were calculated to

be 0.14-0.16 Å longer than was experimentally observed.24

. It might be that addition of counter-

ions further stabilizes Structure 2 enough to reach quantitative agreement with experimental

observation.11

Computed local minimum structures for the [UO2(H2O)F4]2-

complex have generally been

structures with the water ligand at about 3.90-4.02 Å away from the uranium atom12

(Structure 1,

Figure 7.2b). However, experimental studies in crystal structures and solution have indicated the

presence of a short U-OH2 bond about 2.48 or 2.11 Å in length respectively.24, 81

The C2v

structure (Structure 2, Figure 7.2c) corresponding to this arrangement has two imaginary

frequencies in both gas and continuum phase DFT calculations. Structure 2 was calculated to

convert to structure 1 or to dissociate to [UO2F4]2-

and an outgoing water molecule with

conversion or dissociation energies of 10-13 and 5.4-6.3 kcal/mol respectively in aqueous

solution, Table 7.1. Carr-Parrinello MD (CPMD) simulations by Bühl et al11

have been used to

show that the addition of two ammonium counter-ions stabilizes the „experimental‟ geometry of

structure 2 by about 2-4 kcal/mol thus possibly justifying the experimental observation.24, 81

It

would therefore appear that the inclusion of counter-ions according to Bühl et al in calculations

on the [UO2(H2O)F4]2-

complex is essential.11

[UO2F5]3-

Docked in the Cavities of the RHCC. The terminal (N- and C-terminal or cavities

one and four respectively) cavities of the RHCC protein are lined at the N-terminal and C-

terminal ends by isoleucine residues. The intermediate cavities two and three are however

Page 275: Relativistic Quantum Chemistry Applied to Actinides

253

bounded at their N- and C-terminal ends by isoleucine and leucine residues respectively (7.1).

These hydrophobic residues are oriented such that their side chains protrude into the axial

channel of the protein, Figures 7.1 and 7.5. Eight residues from each monomer unit which are

amino acids less hydrophobic than leucine and isoleucine form a ring around each cavity.5 The

cavities range in size from 140-280 Å3 in the native protein and are occupied by water clusters.

In this work, [UO2F5]3-

and cisplatin were docked in these cavities and optimized using

QM/MM. Only preliminary docking simulations were carried out for [Pt(CN)6]2-

. [UO2F5]3-

was

used as a representative for the [UO2Fn(H2O)5-n]2-n

complexes due to the rigidity of the bonds

between the uranyl group and the equatorial ligands and the resulting relative ease to dock in the

hydrophobic cavities. This is in spite of the fact that [UO2F5]3-

is formed only at high fluoride

concentrations.24

[UO2F5]3-

can be thought of as a model system with the protein environment

having similar structural and electronic effects on the other members of the pentaaquo-fluoro

series.

An infinite number of possible orientations of the uranyl group can be expected if

[UO2F5]3-

is embedded in a chemically uniform and spherical hydrophobic cavity. However,

docking of this complex in all the RHCC cavities indicates the presence of two favored

orientations of the uranyl group with respect to the axial channel. The uranyl group can only be

oriented along the two axes perpendicular to the RHCC axis in all the hydrophobic cavities,

Table 7.3. The equatorial fluoro ligands occupy the second axis orthogonal to the axial channel

and no possible alignment of the O=U=O group along the RHCC axial channel was found.

The presence of these minima orientations of the embedded complex along axes

orthogonal to the axial channel can be explained by two observations. Firstly, the cavities of the

RHCC protein are chemically heterogeneous and roughly cylindrical cavities with isoleucine and

Page 276: Relativistic Quantum Chemistry Applied to Actinides

254

or leucine residues (ah-layer) pointing into the center of the helix while the backbone of the

arginine and tyrosine residues (de-layer) form a ring bounding the cavity.5, 10

The uranyl complex

embedded in the cavities is always positioned closest to the isoleucine residues at the N-terminal

end of the cavities except in cavity four, the C-terminal cavity, in which the complex is

positioned nearest to the residues at the C-terminal end due to exposure to the aqueous

environment. Maximum separation from the leucine residues at the C-terminal end of the cavities

could be explained based on steric considerations. The leucine residues simply penetrate deeper

into the axial channel affording lesser space for the embedded ligand. On the other hand, the

isoleucine side chains at the N-terminal ends of the cavities are more „open‟ and have greater

access into the axial channel, Figure 7.1, and the aqueous environment in the case of cavity one,

the N-terminal cavity.

Secondly, the side-chains of the isoleucine residues pointing into the axial channel from

each monomer backbone form a cross-like space bounded at the four edges by the protein

backbone, Figure 7.5. It is near this cross-like opening in the isoleucine residues partitioning the

axial channel that the two possible arrangements of the uranyl moiety and the five equatorial

fluoride ligands of the [UO2F5]3-

complex fit. Typical distances between the alkyl-hydrogen

atoms of the isoleucine residues and the atoms of the [UO2F5]3-

complex ranged from 2.0 to 4.8

Å after the docking.

QM/MM Calculations on [UO2F5]3-

Embedded in the RHCC Cavities. Structurally, the D5H

symmetry of the [UO2F5]3-

complex is lifted and five different U-F bond lengths were observed

after QM/MM optimization in all the cavities. For each cavity, the orientation with longer U-F

bonds or shorter distance between its fluoro ligand and the hydrogen atoms of the isoleucine

residues is less favorable energetically, Table 7.3. Electronically, there is some correspondence

Page 277: Relativistic Quantum Chemistry Applied to Actinides

255

between the U-F bond lengths and the calculated HOMO-LUMO gaps of the embedded ligands.

After considering energy constraints and the calculated structural parameters, it can be concluded

that there is a possibility of experimentally observing the two ligand orientations in only cavity

two.

Figure 7.5: The Atoms of the Isoleucine Residues (Represented as Green Crosses) at the N-

Terminal End of Cavity Two. The Protruding Side-Chains Form a Cross-Like Space into which

the Uranyl Group and Equatorial Fluoride Ligands of [UO2F5]3-

are Embedded. The Two White

Arrows Depict the two Possible Alignments of the Uranyl Moiety and the Equatorial Ligands.

The maximum energy difference between the two orientations in the four cavities

obtained using the docking procedure, ΔΔGDocking, is 0.03 kcal/mol, Table 7.3. The orientations

are therefore essentially iso-energetic and the magnitude of this energy barrier suggests that both

orientations should be experimentally observed in all the cavities with a high probability of inter-

conversion. However, the calculated energy barriers,84

ΔΔGQM/MM, Table 7.3, indicate that the

two orientations of the embedded QM/MM optimized [UO2F5]3-

complex should be observable

only in cavity two, the largest cavity. Only in this cavity are they still iso-energetic, even at the

Page 278: Relativistic Quantum Chemistry Applied to Actinides

256

Tables 7.3: Calculated relative energies (kcal/mol), frontier gaps (eV) and structural features of

the two orientations of the [UO2F5]3-

complex in the hydrophobic cavities of tetrabrachion.

a The two alignments of [UO2F5]

3- in the cavities are labeled „a‟ and „b‟ according to their

relative energies. b The free energies, ΔΔG are given as the difference, ΔGA-ΔGB.

Cav. Allign.a ΔΔG

Dockingb

ΔΔG

QM/MM

Gap

(eV)

RU=O

(Å)

RU-F

(Å)

<UO2

(º)

1 a 2.709 1.877-1.892 2.272-2.372 177.2

b 0.00 10.12 2.634 1.874-1.877 2.273-2.420 176.8

2 a 2.677 1.865-1.876 2.321-2.348 179.2

b 0.03 0.05 2.702 1.878-1.894 2.324-2.355 178.9

3 a 2.489 1.864-1.887 2.266-2.391 175.8

b 0.03 5.77 2.366 1.860-1.892 2.260-2.405 172.1

4 a 2.794 1.877-1.909 2.268-2.391 177.9

b 0.01 8.33 2.793 1.854-1.866 2.214-2.365 177.5

Gas 2.694 1.878 2.337 180.0

Aqueous 2.728 1.871 2.294 180.0

Page 279: Relativistic Quantum Chemistry Applied to Actinides

257

QM/MM level. There is strong experimental evidence for the existence of the two orientations of

uranyl fluoride in this cavity as found by the calculations.53

The other cavities all have one high-

energy conformation resulting in a low probability of experimental observation.

The calculation of the free energies associated with the replacement of the water clusters

in the native RHCC cavities by the [UO2F5]3-

ligand is difficult and requires averaging over the

configuration space of the protein-ligand system. This is particularly important as the water

clusters move on much faster timescales than the uranyl complex. Using a few snapshots of the

protein-ligand system in calculating the free energies associated with displacement of the native

water clusters by [UO2F5]3-

may therefore be misleading. However intuitively, the exchange of

the water clusters with uranyl fluorides would be expected to be exergonic.4-5

The free energy

associated with the displacement of the water clusters could be expected to be dominated by a

large entropic contribution6 even though the enthalpy contribution is expected to be slightly

positive due to the transfer of the uranyl complex from the aqueous medium into a cavity

surrounded by hydrophobic residues.

Both orientations in cavity one, the N-terminal cavity, have U=O and U-F bond lengths

within 0.02 Å and 0.08 Å of the gas-phase structural parameters respectively, Table 7.3. An

examination of the electronic structure of the embedded ligands reveals a reduction of the

HOMO-LUMO gap by approximately 0.06 eV in the less energetically favored orientation. This

orientation also has the larger U-F bond deformation and a greater change in the O=U=O bond

angle.

The structural parameters of the embedded ligands in cavity two (the largest cavity) are

essentially identical to those calculated for the gas-phase structure. The U=O and U-F bond

Page 280: Relativistic Quantum Chemistry Applied to Actinides

258

lengths are all within 0.03 Å of the calculated gas-phase values. In addition, the uranyl bond

angles are within 1.2º of the gas-phase structures. The similarity of the ligands embedded in this

cavity to the gas-phase complex is also reflected in the calculated uranyl stretching vibrational

frequencies. Electronically, the HOMO-LUMO gaps and the description of the frontier orbitals

for the ligand embedded in this cavity are similar (to within about 0.02 eV) to those calculated

for the gas phase structure. The similarity of both uranyl conformations to the gas-phase

structure is most likely due to the size of the cavity, the large size of the cavity allowing for

minimal interaction with the hydrogen atoms of the isoleucine residues.

The effects of the RHCC framework on the geometries of the embedded ligands are more

pronounced in cavity three than in the first two cavities. Deviations of the U=O and U-F bond

lengths by 0.03 Å and 0.07 Å respectively from the gas-phase values were calculated in cavity

three. The uranyl bond angle was decreased by 4.2 and 7.9º in the two possible orientations in

this cavity. A significant decrease in the HOMO-LUMO gap by 0.21-0.33 eV was calculated in

the ligands embedded in this cavity compared to the gas phase complex.

The less energetically favorable orientation in cavity four, the C-terminal cavity, has

significant deformation of the U-F bonds by up to 0.12 Å. There is lesser deformation of the U-F

bonds (0.02-0.07 Å) and longer U=O bonds in the more energetically accessible orientation,

Table 7.3. The U=O bonds in both orientations are however within 0.03 Å of the gas-phase

structure. In addition, the HOMO-LUMO gaps for both orientations of the ligand are similar and

are about 0.1 eV larger than the gas-phase value.

Other Ligands Embedded in the RHCC. Due to the potential use of the cavities of this coiled-

coil protein as a delivery vehicle for therapeutic molecules, we have also explored the enclosure

Page 281: Relativistic Quantum Chemistry Applied to Actinides

259

of a cisplatin molecule in cavity two of the RHCC. Unlike in the case of [UO2F5]3-

, docking and

subsequent QM/MM calculations of cisplatin in this cavity indicate that there are more than two

possible orientations with respect to the axial channel. Also, there is no preference for alignment

or location towards the isoleucine residues at the N-terminal end of the cavity unlike the uranyl

complex.

An explanation for this is the absence of a strong oxo- axial group that can preferentially

anchor towards the isoleucine residues at the N-terminal end of the cavity. The chloro- and

ammine ligands of cisplatin can therefore be oriented in a fairly large number of possible

orientations with respect to the RHCC axis. The chemical nature of the ammine ligands also

allows for interaction with the carbonyl groups of the cavity wall. On the other hand, preliminary

docking of the hexacyano platinate complex, [Pt(CN)6]2-

in the largest cavity revealed only two

orientations reminiscent of those observed for the uranyl complex. The presence of two favored

orientations for [UO2F5]3-

and [Pt(CN)6]2-

in the protein cavities suggests that these alignments

will exist for molecules with strong axial groups in an octahedral or pentagonal bipyramidal

framework.

The calculated structural parameters for two randomly selected poses of cisplatin

embedded in the largest cavity of the RHCC protein are compared to the calculated gas and

aqueous-phase structures as well as the experimental crystal structure85

in Table 7.4. Overall the

optimized structures of these two poses are essentially similar to the gas phase structure even

though there is a minute elongation of the Pt-Cl bond as well as contraction of the N-Pt-N bond

angle. This is largely not surprising as very few water molecules actually penetrate the axial

channel of the RHCC protein and cavity two is large enough to allow the molecule to exist in a

„pseudo-gaseous‟ state.

Page 282: Relativistic Quantum Chemistry Applied to Actinides

260

Table 7.4: Computed Structural Parameters Cisplatin in the Gaseous and Aqueous Phases and of

Two Randomly Selected Orientations of Cisplatin Embedded in the Cavity Two (Largest Cavity)

of the RHCC Protein Obtained Using RECPs.

Parameter Gaseous Solution Expt.* Structure-1 Structure-2

B3LYP BP86 B3LYP BP86 BP86 BP86

Pt-Cl 2.310 2.303 2.349 2.336 2.330 2.312 2.314

Pt-N 2.104 2.092 2.079 2.068 2.010 2.099 2.101

N-Pt-N 98.19 99.27 92.22 92.99 87.0 98.81 98.26

Cl-Pt-Cl 95.54 95.78 93.92 94.10 91.9 95.53 95.63

* Crystal Structure from Reference 82

.

Conclusions

A systematic study of pentacoordinated aquo and fluoro uranyl complexes has been

carried out using two different relativistic methods, the relativistic effective core potentials and

ZORA with all electron basis set, in conjunction with the B3LYP and BP86 density functionals.

The effects of an aqueous medium on the geometrical structure, ligand binding energies and

electronic structure were investigated using the COSMO solvation model as well as an explicit

second coordination sphere. The conformational alignments, electronic structures and

geometrical parameters of [UO2F5]3-

and cisplatin embedded in the hydrophobic cavities of the

tetrabrachion coiled coil have been determined with the hybrid QM/MM approach. The studies

of cisplatin and [UO2F5]3-

inside the hydrophobic cavities were motivated by a need for insight

Page 283: Relativistic Quantum Chemistry Applied to Actinides

261

into therapeutic usage of the protein in drug delivery and structural or conformational alignment

reasons, respectively.

The inclusion of solvation effects using the COSMO model generally leads to only slight

increases in the U=O bonds lengths in contrast to a contraction of the U-F bonds. This effect is

associated with decreases in the uranyl stretching vibrational frequencies. The calculated U=O

bond lengths increase as the number of fluoride ligands are increased in both the gas and aqueous

phases. The range of the U=O bond lengths from [UO2(H2O)5]2+

to [UO2F5]3-

is however only

accurately captured after the inclusion of five water molecules in the second coordination sphere.

The inclusion of water molecules in the second coordination sphere also results in better

agreement between the calculated and experimental uranyl vibrational frequencies.

Although, the docking procedure used in this contribution indicates the presence of two

favored orientations of the [UO2F5]3-

complex in all four hydrophobic cavities of the

tetrabrachion coiled-coil protein, the calculated relative free energies of embedded complexes

optimized at the QM/MM level however reveal that both orientations can only be experimentally

observed in cavity two, the largest cavity. There is strong experimental evidence for these

orientations as found by the calculations.53

In the other three cavities, only one conformation is

energetically accessible. The two possible orientations of the uranyl complexes in the protein

cavities are along the two axes perpendicular to the protein channel axis. There is no possible

orientation of the embedded ligands along the channel axis. The uranyl pentafluoride is generally

associated with the isoleucine residues at the N-terminus end of the protein cavities. This is due

to a combination of steric effects and interaction with the alkyl side chains of the isoleucine

residues. The presence of two ordered orientations of the uranyl complexes in the hydrophobic

cavities is not unique. Docking of the hexacoordinated [Pt(CN)6]2-

complex also reveals a similar

Page 284: Relativistic Quantum Chemistry Applied to Actinides

262

structure in cavity two of the protein. On the other hand, the anticancer drug, cisplatin, shows no

preferred orientations in the protein cavities.

In general, there is little change in the U=O bond lengths upon embedding of the

complexes, with a maximum change of 0.03 Å from the gas phase value of 1.878 Å at the

BP86/RECP level. The largest structural changes are seen in the U-F bonds and O=U=O bond

angles. An examination of the structural features, HOMO-LUMO gaps and uranyl stretching

vibrational frequencies of the uranyl complexes embedded in cavity two reveal great similarity to

the gas phase structure.

References

1. Burkhard, P.; Stetefeld, J.; Strelkov, S. V., Trends Cell Biol. 2001, 11, 82.

2. Eriksson, M.; Hassan, S.; Larsson, R.; Linder, S.; Ramqvist, T.; Lovborg, H.; Vikinge,

T.; Figgemeier, E.; Muller, J.; Stetefeld, J.; Dalianis, T.; Ozbek, S., Anticancer Res. 2009, 29,

11.

3. http://www.pdb.org/pdb/explore/explore.do?structureId=1FE6.

4. Ozbek, S.; Muller, J. F.; Figgemeier, E.; Stetefeld, J., Acta Crystallogr. D. 2005, 61, 477.

5. Stetefeld, J.; Jenny, M.; Schulthess, T.; Landwehr, R.; Engel, J.; Kammerer, R. A., Nat.

Struct. Biol. 2000, 7, 772.

6. Yin, H.; Hummer, G.; Rasaiah, J. C., J. Am. Chem. Soc. 2007, 129, 7369.

7. Yin, H.; Hummer, G.; Rasaiah, J. C., JACS 2007, 129, 7369.

8. McFarlane, A. A.; Orriss, G. L.; Stetefeld, J., Eur. J. Pharmacol. 2009, 625, 101.

9. Ozbek, S.; Muller, J. F.; Figgemeier, E.; Stetefeld, J., Acta Crystallogr., Sect. D: Biol.

Crystallogr. 2005, 61, 477.

Page 285: Relativistic Quantum Chemistry Applied to Actinides

263

10. Stetefeld, J.; Jenny, M.; Schulthess, T.; Landwehr, R.; Engel, J.; Kammerer, R. A., Nat.

Struct. Biol. 2000, 7, 772.

11. Bühl, M.; Schreckenbach, G.; Sieffert, N.; Wipff, G., Inorg. Chem. 2009, 48, 9977.

12. Bühl, M.; Sieffert, N.; Wipff, G., Chem. Phys. Lett. 2009, 467, 287.

13. Gaillard, C.; El Azzi, A.; Billard, I.; Bolvin, H.; Hennig, C., Inorg. Chem. 2005, 44, 852.

14. Garcia-Hernandez, M.; Willnauer, C.; Krüger, S.; Moskaleva, L. V.; Rösch, N., Inorg.

Chem. 2006, 45, 1356.

15. Infante, I.; Van Stralen, B.; Visscher, L., J. Comput. Chem 2006, 27, 1156.

16. Infante, I.; Visscher, L., J. Comput. Chem 2004, 25, 386.

17. Macak, P.; Tsushima, S.; Wahlgren, U.; Grenthe, I., Dalton T. 2006, 3638.

18. Paez-Hernandez, D.; Ramirez-Tagle, R.; Codorniu-Hernandez, E.; Montero-Cabrera, L.

A.; Arratia-Perez, R., Polyhedron 2010, 29, 975.

19. Schreckenbach, G., Inorg. Chem. 2002, 41, 6560.

20. Schreckenbach, G.; Hay, P. J.; Martin, R. L., J. Comput. Chem 1999, 20, 70.

21. Shamov, G. A.; Schreckenbach, G.; Vo, T. N., Chem-Eur. J. 2007, 13, 4932.

22. Straka, M.; Dyall, K. G.; Pyykkö, P., Theor. Chem. Acc. 2001, 106, 393.

23. Straka, M.; Kaupp, M., Chem. Phys. 2005, 311, 45.

24. Vallet, V.; Wahlgren, U.; Schimmelpfennig, B.; Moll, H.; Szabo, Z.; Grenthe, I., Inorg.

Chem. 2001, 40, 3516.

25. Wang, Q.; Pitzer, R. M., J. Phys. Chem. A. 2001, 105, 8370.

26. Zachariasen, W. H., Acta Crystallogr. 1954, 7, 783.

27. Cramer, C. J., Essentials of Computational Chemistry: Theories and Models. 2nd ed.;

John Wiley & Sons Ltd: West Sussex, 2004; p 17.

Page 286: Relativistic Quantum Chemistry Applied to Actinides

264

28. Eichinger, M.; Tavan, P.; Hutter, J.; Parrinello, M., J. Chem. Phys. 1999, 110, 10452.

29. Field, M. J., J. Comput. Chem 2002, 23, 48.

30. Friesner, R. A.; Guallar, V., Ann. Rev. Phys. Chem. 2005, 56, 389.

31. Hu, H.; Elstner, M.; Hermans, J., Proteins 2003, 50, 451.

32. Lin, H.; Truhlar, D. G., Theor. Chem. Acc. 2007, 117, 185.

33. Murphy, R. B.; Philipp, D. M.; Friesner, R. A., J. Comput. Chem 2000, 21, 1442.

34. Schoneboom, J. C.; Lin, H.; Reuter, N.; Thiel, W.; Cohen, S.; Ogliaro, F.; Shaik, S., J.

Am. Chem. Soc. 2002, 124, 8142.

35. Senn, H. M.; Thiel, W., QM/MM methods for biological systems. In Atomistic

Approaches in Modern Biology: From Quantum Chemistry to Molecular Simulations, 2007; Vol.

268, pp 173.

36. Sherwood, P.; de Vries, A. H.; Guest, M. F.; Schreckenbach, G.; Catlow, C. R. A.;

French, S. A.; Sokol, A. A.; Bromley, S. T.; Thiel, W.; Turner, A. J.; Billeter, S.; Terstegen, F.;

Thiel, S.; Kendrick, J.; Rogers, S. C.; Casci, J.; Watson, M.; King, F.; Karlsen, E.; Sjovoll, M.;

Fahmi, A.; Schafer, A.; Lennartz, C., J. Mol. Struc-Theochem. 2003, 632, 1.

37. Vreven, T.; Morokuma, K.; Farkas, O.; Schlegel, H. B.; Frisch, M. J., J. Comput. Chem

2003, 24, 760.

38. Warshel, A., Ann. Rev. Bioph. Biom. 2003, 32, 425.

39. Woo, T. K.; Cavallo, L.; Ziegler, T., Theor. Chem. Acc. 1998, 100, 307.

40. Warshel, A.; Levitt, M., J. Mol Biol. 1976, 103, 227.

41. Singh, U. C.; Kollman, P. A., J. Comput. Chem 1986, 7, 718.

42. Valiev, M.; Yang, J.; Adams, J. A.; S., T. S.; Weare, J. H., J. Phys. Chem. B 2007, 111,

13455.

Page 287: Relativistic Quantum Chemistry Applied to Actinides

265

43. Friesner, R. A.; Guallar, V., Annu. Rev. Phys. Chem. 2005, 56, 389.

44. Sauer, J.; Sierka, M., J. Comput. Chem. 2000, 21, 1470.

45. Senn, H. M.; Thiel, W., Atomistic Approaches in Modern Biology: From Quantum

Chemistry to Molecular Simulations 2007, 268, 173.

46. Shurki, A.; Warshel, A., Protein Simulations 2003, 66, 249.

47. Infante, I.; Van Stralen, B.; Visscher, L., J. Comput. Chem. 2006, 27, 1156.

48. Infante, I.; Visscher, L., J. Comput. Chem. 2004, 25, 386.

49. Faas, S.; Snijders, J. G.; van Lenthe, J. H.; van Lenthe, E.; Baerends, E. J., Chem. Phys.

Lett. 1995, 246, 632.

50. van Lenthe, E., J. Comp. Chem. 1999, 20, 51.

51. van Lenthe, E.; Baerends, E. J.; Snijders, J. G., J. Chem. Phys. 1993, 99, 4597.

52. Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.; Ferguson, D. M.;

Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.; Kollman, P. A., J. Am. Chem. Soc. 1995, 117, 5179.

53. Stetefeld, J.; Meier, M., Unpublished Results.

54. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J., J. Phys. Chem. 1994, 98,

11623.

55. Becke, A. D., J. Chem. Phys. 1993, 98, 5648.

56. Lee, C. T.; Yang, W. T.; Parr, R. G., Phys. Rev. B 1988, 37, 785.

57. Becke, A. D., Phys. Rev. A 1988, 38, 3098.

58. Perdew, J. P., Phys. Rev. B 1986, 33, 8822.

59. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100, 7535.

60. Cao, X. Y.; Moritz, A.; Dolg, M., Chem. Phys. 2008, 343, 250.

61. http://www.theochem.uni-stuttgart.de/pseudopotentials/clickpse.en.html.

Page 288: Relativistic Quantum Chemistry Applied to Actinides

266

62. Bylaska, E. J. d. J., W. A.; Govind N.; Kowalski K.; Straatsma, T. P.; Valiev, M.; Wang,

D.; Apra, E.; Windus, T. L.; Hammond, J.; Nichols, P.; Hirata, S.; Hackler, M. T.; Zhao, Y.; Fan,

P. -D.; Harrison, R. J.; Dupuis, M.; Smith, D. M. A.; Nieplocha, J.; Tipparaju, V.; Krishnan, M.;

Wu, Q.; Van Voorhis, T.; Auer, A. A.; Nooijen, M.; Brown, E.; Cisneros, G.; Fann, G. I.;

Fruchtl, H.; Garza, J.; Hirao, K.; Kendall, R.; Nichols, J. A.; Tsemekhman, K.; Wolinski, K.;

Anchell, J.; Bernholdt, D.; Borowski, P.; Clark, T.; Clerc, D.; Dachsel, H.; Deegan, M.; Dyall,

K.; Elwood, D.; Glendening, E.; Gutowski, M.; Hess, A.; Jaffe, J.; Johnson, B.; Ju, J.;

Kobayashi, R.; Kutteh, R.; Lin, Z.; Littlefield, R.; Long, X.; Meng, B.; Nakajima, T.; Niu, S.;

Pollack, L.; Rosing, M.; Sandrone, G.; Stave, M.; Taylor, H.; Thomas, G.; van Lenthe, J.; Wong,

A.; and Zhang, Z. NWChem, A Computational Chemistry Package for Parallel Computers,

Version 5.1". 2007.

63. Valiev, V.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T. P.; van Dam, H. J. J.;

Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L.; de Jong, W. A., Comput. Phys. Commun.

2010, 181, 1477.

64. Klamt, A., J. Phys. Chem. 1995, 99, 2224.

65. Klamt, A.; Schuurmann, G., J. Chem. Soc. Perk T. 2 1993, 799.

66. te Velde, G.; Bickelhaupt, F. M.; van Gisbergen, S. J. A.; Fonseca Guerra, C.; Baerends,

E. J.; Snijders, J. G.; Ziegler, T., J. Comput. Chem. 2001, 22, 931.

67. Fonseca Guerra, C.; Snijders, J. G.; te Velde, G.; Baerends, E. J., Theor. Chem. Acc.

1998, 99, 391.

68. ADF2007.01, Theoretical Chemistry, Vrije Universiteit, Amsterdam, The Netherlands,

http://www.scm.com.

69. Swart, M.; van Duijnen, P. T.; Snijders, J. G., J Comput Chem 2001, 22, 79.

Page 289: Relativistic Quantum Chemistry Applied to Actinides

267

70. Pye, C. C.; Ziegler, T., Theor. Chem. Acc. 1999, 101, 396.

71. Schreckenbach, G.; Shamov, G. A., Accounts Chem. Res.search 2009, in print

10.1021/ar800271r.

72. Goodsell, D. S.; Morris, G. M.; Olson, A. J., J. Mol. Recognit. 1996, 9, 1.

73. Guilbaud, P.; Wipff, G., J. Mol. Struct. THEOCHEM 1996, 366, 55.

74. Spiegel, K.; Rothlisberger, U.; Carloni, P., J. Phys. Chem. B 2004, 108, 2699.

75. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2010, 114, 1957.

76. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109, 10961.

77. Schultz, N. E.; Zhao, Y.; Truhlar, D. G., J. Phys. Chem. A. 2005, 109, 4388.

78. Vallet, V.; Wahlgren, U.; Szabo, Z.; Grenthe, I., Inorg. Chem. 2002, 41, 5626.

79. Vallet, V.; Wahlgren, U.; Grenthe, I., J. Am. Chem. Soc. 2003, 125, 14941.

80. Flint, C. D.; Tanner, P. A., Mol. Phys. 1981, 43, 933.

81. Mak, T. C. W.; Yip, W. H., Inorg. Chim. Acta 1985, 109, 131.

82. Jones, L. H.; Penneman, R. A., J. Chem. Phys. 1953, 21, 542.

83. Perron, H.; Roques, J.; Domain, C.; Drot, R.; Simoni, E.; Catalette, H., Inorg. Chem.

2008, 47, 10991.

84. Valiev, M.; Garrett, B. C.; Tsai, M. K.; Kowalski, K.; Kathmann, S. M.; Schenter, G. K.;

Dupuis, M., J. Chem. Phys. 2007, 127.

85. Milburn, G. H. W.; Truter, M. R., J. Chem. Soc. A 1966, 1, 1609.

86. Guillaumont, R.; Fanghänel, T.; Grenthe, I.; Neck, V.; Palmer, D. A.; Rand, M. H.,

Update on the Chemical Thermodynamics of Uranium, Neptunium, Plutonium, Americium and

Technetium, . Elsevier: Amsterdam, 2003; Vol. 5, p 970.

Page 290: Relativistic Quantum Chemistry Applied to Actinides

268

Preface to Chapter 8

This chapter is based on two manuscripts. The full citations of these manuscripts are:

1) P. L. Arnold, G. M. Jones, S. O. Odoh, G. Schreckenbach and J. B. Love. “Strongly

coupled binuclear uranium-oxo complexes from uranyl oxo rearrangement and

reductive silylation”. Nature Chemistry, 2012, 4, 221.

2) P. L. Arnold, E. Holis, J. B. Love, N. Magnani, E. Colineau, R. Caciuffo, N.

Edelstein, L. Castro, A. Yahia, L. Maron, S. O. Odoh and G. Schreckenbach.. “Oxo-

Functionalization and Reduction of the Uranyl Ion through Lanthanide-Element Bond

Homolysis; Synthetic, Structural, and Bonding Analysis of a Series of Singly

Reduced Uranyl - Rare Earth 5f1-4f

n Complexes”. 2012, to be submitted.

In this chapter, the results of our theoretical calculations on two types of complexes synthesized

and characterized by our experimental collaborators are presented. In the first manuscript, a bis-

oxo-silyl uranium (V) complex with a butterfly shaped U2O4 core was characterized. This novel

U2O4 core is unique as a result of the fact that one of the UO2 groups was transformed from the

traditionally linear format (as found in the starting material) into a bent/cis orientation. In

contrast, in the second manuscript two complexes with U2O4 cores in which the UO2 groups

interact by side-on cation-cation interactions were characterized. The two classes of complexes

studied in this work are stable U(V)/U(V) bi-nuclear complexes in which reduction from the

hexavalent state was achieved either by oxo-silylation (in the case of the butterfly-shaped

complex) and oxo-metallation by transition metals (in the case of the side-on cation-cation

complexes).

Page 291: Relativistic Quantum Chemistry Applied to Actinides

269

In the first manuscript, all theoretical DFT calculations of the structure and electronic properties

of the bis-nuclear butterfly-type complex were carried out by Samuel O. Odoh. The second

manuscript contains two sets of calculations: a) a B3LYP/RECP approach was employed by the

Maron group from France and b) the all-electron approach with the PBE functional was used in

the calculations carried out by me as part of the contributions from the Schreckenbach group in

Manitoba, Canada.

Although this chapter is based on our contributions to the two manuscripts, it has been

completely re-written to emphasize our contributions. Our studies of the two types of complexes

were deliberately collated in one chapter due to the similarity of a U(V)/U(V) framework as well

as reduction from the hexavalent to pentavalent state facilitated by oxo-functionalization.

Copyright permissions have been obtained from the publishers of this article and from all the

other authors. Polly L. Arnold, Emmalina Hollis, G.M. Jones, Gary S. Nichols and Jason B. Love

are affiliated with the School of Chemistry, University of Edinburgh, Edinburgh, United

Kingdom. Jean-Christophe Griveau, Roberto Caciuffo

are affiliated with the European

Commission, Joint research centre, Institute for Transuranium Elements, Karlsruhe, Germany.

Nicola Magnani and Norman Edelstein are affiliated with the Lawrence Berkeley National

Laboratory, Berkeley, U.S.A. Laurent Maron, Ludovic Castro and Ahmed Yahia are affiliated

with the University of Toulouse, Toulouse, France.

Page 292: Relativistic Quantum Chemistry Applied to Actinides

270

Chapter 8: Bis-Uranium (V) Dioxo Complexes formed after Oxo-

functionalization of Uranyl Axial Oxo Atoms

Abstract

Bis-uranium (V) dioxo complexes of the Pacman-type macrocycle have been characterized using

scalar relativistic density functional theory. The class of bimetallic complexes studied included

cases in which the U2O4 unit is aggregated in either a butterfly-shaped fashion or via cation-

cation interactions between the pentavalent uranyl groups in a diamond-shaped motif. The axial

oxo atoms of the butterfly-shaped moiety are functionalized by silyl groups while transition

metals, yttrium and samarium, were used in reductively functionalization of the axial oxo atoms

in the diamond-shaped complexes. The formation of the butterfly-shaped motif most likely

proceeds via a pentavalent intermediate rather than via a hexavalent complex. For the complexes

with a diamond-shaped motif, the calculated structural parameters agree well with the

experimental data. Accurate prediction of the energetic ordering of the low-lying electronic

states in these complexes appear to be problematic for the methods used in this study.

Introduction

There has been a significant resurgence in the synthesis and characterization of various

uranium complexes with organic ligands such as macrocycles and expanded porphyrins.1-11

The

pentavalent state of the uranyl moiety is however difficult to isolate and characterize in an

aqueous environment due to its susceptibility to disproportionation and oxidation. Ikeda and co-

workers used a spectro-electrochemical approach to study uranyl pentavalent complexes by

Page 293: Relativistic Quantum Chemistry Applied to Actinides

271

electrochemically reducing hexavalent complexes.12-16

The properties of the quasi-stable

pentavalent complexes produced in the optically transparent electrochemical cell were then

measured using spectroscopic approaches such as nuclear magnetic resonance (NMR)13

, infra-

red (IR)15

and extended X-ray absorption fine structure (EXAFS)12

. The capability to synthesize

and isolate pentavalent UO2 complexes has increased greatly since the isolation of a pentavalent

uranyl triphenyl phosphine oxide cation with a triflate anion pair by Berthet et al.17

More

recently, an approach involving reductive oxo-functionalization of the oxo- atoms of UO22+

complexes has been used by a number of workers.3, 18-21

Complexes in which one or both oxo-

atoms of the uranyl have been functionalized with strong Lewis acids such as silyl groups19-20

,

hydrogen atoms20

and B(C6F5)322

have all been reported.

Discrete complexes in which an oxo- atom(s) of a uranyl group interacts with the

uranium atom of another uranyl group have also been synthesized.6, 23-27

Cation-cation

complexes involving uranyl-uranyl interactions constitute a significant fraction of kinetically

stable pentavalent uranyl complexes.3 There are generally two structural motifs for cation-cation

interactions in multinuclear uranyl complexes, the diamond motif in which the two uranyl groups

are in (nearly) parallel arrangement22

and the aptly named T- shaped motif, in which the uranyl

groups are (nearly) perpendicular, Figure 8.1. An example of note is the tetrameric cation-cation

complex reported by Burdet et al.23

The adjacent uranyl groups interact through a T-shaped

cation-cation motif, Figure 8.1.

Page 294: Relativistic Quantum Chemistry Applied to Actinides

272

Figure 8.1: Two common motifs for cation-cation interactions (CCI) between uranyl groups. (A)

The diamond motif in which the UO2 groups are arranged in a parallel fashion. (B) The T-shaped

motif in which an oxo- atom of one uranyl group interacts with the uranium atom of a

neighboring uranyl group. All oxygen, uranium and carbon atoms in this work are depicted as

red, blue and grey colors respectively.

In the first part of this chapter the structural and electronic properties of a recently

synthesized and characterized complex, 1, Figure 8.2, featuring a butterfly-shaped U2O4 motif is

presented.28

The butterfly-shaped U2O4 motif is different from the common uranyl-uranyl CCI

motifs shown in Figure 8.1. In this complex, the U2O4 core is encased in a Pacman-type, H4L

ligand, Figure 8.2 with the two exo- oxo atoms reductively functionalized by silyl groups. There

is a particular feature of the binuclear butterfly-shape that is worthy of note. The orientation of

the Oexo-U-Ocis moiety implies that one of the traditionally linear UO2 groups has been bent into a

cis/bent dioxouranium moiety.

Page 295: Relativistic Quantum Chemistry Applied to Actinides

273

Figure 8.2: The recently synthesized U2O4 Pacman complex, 1a. (Left) A side-on view of

complex 1a. The U2O4 core in this complex is butterfly-shaped. O1 and O4 are the exo- oxo

atoms while O2 and O3 are the endo- and cis- oxo atoms respectively. The top view of the

Pacman-type H4L ligand is shown on the right. The two amine sites of this ligand are occupied

by U(V) centers in 1a.

Reductive oxo-functionalization of the axial oxo atoms of the UO22+

moiety has been

achieved with silyl-type groups as in 1, metal ions and other groups.1, 3, 18-20, 22, 29

Arnold et al.

have reported the synthesis and characterization of several uranyl Pacman complexes in which

the axial oxo atoms have been reductively functionalized by manganese, iron and other transition

metals ions.18

Theoretical characterization of these species was carried out by Berard et al.30

These complexes can be labeled as XOUO(py)H2L species with X being either the proton,

methyl, silyl or the metal ion. The equatorial coordination of the dioxouranium (V) units in these

species is satisfied by two linkages to the amine groups of the Pacman-type ligand, two to the

pyrrole groups of the Pacman-type ligand, and a pyridine solvent ligand. Arnold et al.31

have

Page 296: Relativistic Quantum Chemistry Applied to Actinides

274

however synthesized and characterized a new class of oxo-functionalized complexes. In their

work, the dimerization of the XOUO(py)H2L complexes (where X is Y or Sm) results in the

formation of (XOUO)2L2 species. In this work, the yttrium salt is labeled as 2 while its samarium

counterpart is labeled as 3. The yttrium salt is shown in Figure 8.3. The dioxouranium units in

these binuclear complexes are arranged in a diamond U2O4 motif, Figure 8.1. During follow-up

studies of their synthesis of 2 and 3, Arnold et al. have been able to isolate and characterize the

monomer precursors, as lithium chloride salts, of these binuclear complexes.32

The geometry of

the lithium chloride monomer salt of the yttrium dimer is shown in Figure 8.4. The monomer

salts are labeled in a similar manner to the corresponding dimer species. The yttrium monomer is

labeled as 2M and its samarium counterpart as 3M. The M is used to signify that these are the

monomeric salts.

Figure 8.3: Optimized structure of the yttrium dimer complex, 2, obtained at the PBE/L1 level.

The O, U and Y atoms are in red, dark green and light green colours respectively.

Page 297: Relativistic Quantum Chemistry Applied to Actinides

275

Figure 8.4: The lithium-chloride monomer salt of the yttrium dimer complex, 2M. The dioxo-

uranium core with Li and Y oxo-functionalization are shown on the right.

In this report, the structural and electronic properties of the novel uranium-Pacman

complexes, 1, 2 and 3 have been studied using density functional theory (DFT) calculations. The

aim is to characterize their structures with particular emphasis on the nature of their U2O4 cores,

their ground electronic states as well as the possibility of interaction between their actinide

centers. The structure and electronic properties of the lithium-chloride monomer salts, 2M and

3M are also examined.32

We have employed DFT in the calculations in this work. Scalar

relativistic DFT is increasingly the favorite approach for computing the properties of actinide

complexes due to its speed, good treatment of electron correlation effects and general accuracy.

The ability of DFT to provide accurate structural properties, redox potentials and vibrational

frequencies for open-shell actinide complexes has been established by various workers.33

Page 298: Relativistic Quantum Chemistry Applied to Actinides

276

Computational Details

The geometries of 1, 2 and 3 as well as the lithiated monomers, 2M and 3M, were

optimized in the gas phase using DFT calculations with the B3LYP functional.34-35

Single point

calculations in the pyridine solvent were carried out on the optimized structures by employing

the polarizable continuum solvation (PCM) model36-37

in the DFT calculations. The uranium

atom was described with the Stuttgart relativistic pseudopotential38-39

while all other atoms were

described with the 6-31G** basis set. These calculations were performed with the Gaussian 0340

suite of programs and are labeled the B3LYP/RECP calculations. Natural bond orbital (NBO)

analyses were also performed.41

This allowed us to calculate the Mayer-Mulliken bond orders

and the Wiberg bond indices.42

Scalar relativistic calculations with all-electron (AE) basis sets

using a four-component approach were also carried out with the Priroda program.43

The PBE

functional was employed in these calculations while using a triple-ζ (cc-pVTZ) basis set. These

calculations were labeled as the PBE/AE/4-component calculations. The small-component

portion was described using appropriate kinetically balanced basis sets. The Mayer bond orders

were calculated after the geometry optimization.42

Results and Discussion

Electronic structure analysis of 1. Analysis of the bonding in the U2O4 core of 1 was

undertaken using density functional theory (DFT) and natural bond order (NBO) calculations.

Single point calculations in a pyridine solvent continuum were carried out on a molecule of 1

whose geometry had been optimised in the gas phase. Three possible arrangements of the two

uranium-centered f-electrons were considered: triplet (ferromagnetically-coupled, fαfα); anti-

ferromagnetic unrestricted broken-symmetry singlet (fαfβ independently-localised orbitals); and

restricted singlet (fαβ

configuration). The anti-ferromagnetic singlet state was calculated to be

Page 299: Relativistic Quantum Chemistry Applied to Actinides

277

more stable than the triplet state by 1.5 and 1.4 kcal/mol in the gas and pyridine solvent phases

when the B3LYP functional was employed with relativistic pseudopotentials. The energy

difference between these states was calculated as 2.8 kcal/mol in all-electron calculations with

the PBE functional. At this level, the restricted singlet is about 16.0 and 14.2 kcal/mol higher in

energy than the unrestricted singlet and triplet states respectively.

The calculated structural properties of 1 obtained at the B3LYP/RECP and PBE/AE/4-

component levels are compiled in Table 8.1. For the unrestricted singlet state of 1, the bonds

between the uranium and the endo- and cis- oxo atoms were calculated to be between 2.092 and

2.099 Å while the U-Oexo bond lengths were calculated as 2.053 Å, within 0.01 Å of those

obtained experimentally. The experimental structural parameters were obtained from X-ray

diffraction studies.28

Overall the best agreement between the calculated and experimental

structural parameters was obtained for the unrestricted antiferromagnetic singlet electronic state,

Table 8.1. The U···U separation was calculated as 3.366 and 3.379 Å in the unrestricted broken-

symmetry singlet and triplet states respectively. The calculated Mayer bond orders for the Oexo-

Si, U-Oexo, U-Oendo and U-Ocis were calculated as 1.04, 1.27, 1.20 and 1.19 respectively. The

bonds within the U2O4 core can therefore be considered formally as single bonds with some

double bond character, although the U-Oexo bonds are slightly stronger. This is the case for both

the electronic triplet and unrestricted singlet states.

The α-(HOMO-27) and β-(HOMO-27) orbitals obtained with the B3LYP functional for

the unrestricted singlet state are depicted in Figure 8.5(a) and (b) and describe the primary σ-

bonding interaction in the U2O4 core. The contributions from the trans-endo-oxo atom to these

orbitals are significantly larger than those from the cis-oxo atom. There is another set of σ-type

orbitals at slightly higher energy with greater contributions from the cis-oxo atom. Thus,

Page 300: Relativistic Quantum Chemistry Applied to Actinides

278

although the σ-framework is weaker in 1a than in the calculated structure of the hexavalent

uranyl analogue 44

, a strongly bound cis-oxo component can be identified in 1a. In addition, there

is a weaker π-type bonding interaction in the U2O2 core, Figure 8.5c and 8.5d. The π-type

orbitals, Figure 8.5(c) and (d), are dominated by 2p-contributions from the cis-oxo atom and

appear to be only remnants of more prominent and stable π-interactions in the calculated

structure of the hexavalent, non-silylated counterpart 44

.

The combined structural and computational data show that the butterfly U2O4Si2 motif,

Figure 8.2, can be formulated as singly bonded uranium oxo and siloxide groups combined with

a significant π-bonding contribution from the cis-oxo group. Formally, this has resulted from the

rearrangement of two linear, pentavalent actinyls into a new bonding mode for uranium in which

one oxo group is shared and trans- and one has adopted a cis-position. The diamond U2O2

geometry adopted in 1 has been observed in related Group 6 chemistry.28

For example, oxidation

reactions of the quadruply metal-metal bonded Mo acetate dimer in the presence of good π-

accepting ligands form MoV(μ-O)2Mo

V complexes which have single M-M bonds. It is therefore

tempting to look for a direct metal-metal interaction in 1 since no f-block metal-metal bonded

complex has been reported. The calculated U∙∙∙U separation of 3.366 Å is particularly short

(twice the covalent radius of the uranium atom = 3.92 Å) which may indicate some bonding

interaction. This is reflected in the non-trivial calculated Mayer bond order of 0.34 between the

between the uranium atoms, which is only slightly lower than those calculated for some of the U-

N bonds in 1 (range 0.38-0.55). NBO analysis allows the identification of a set of bonding and

antibonding orbitals, Figure 8.6(a) and (b) respectively, between the uranium centers. While

there are no reported examples of molecular bonds between two f-block elements, multiple U-U

bonds with distinctly different interactions compared to transition metals have been predicted by

Page 301: Relativistic Quantum Chemistry Applied to Actinides

279

Table 8.1: Calculated and experimental structural parameters of 1 in the ferromagnetic triplet

and antiferromagnetic broken-symmetry electronic states. The properties obtained for the

antiferromagnetic state are enclosed in parentheses.

B3LYP/RECP PBE/AE/4-Component Expt.

Bond Lengths

(Å)

Mayer BO Bond Lengths

(Å)

Mayer BO

fαfα (fαfβ) (fαfβ) fαfα (fαfβ) fαfα (fαfβ)

U-U 3.379 (3.366) 0.219 3.372 (3.372) 0.34 (0.34) 3.3556(5)

U-Oexo 2.056 (2.053) 0.991 2.055 (2.050) 1.26 (1.27) 2.041(6)

U-Oendo 2.048, (2.092)

2.048, (2.099)

2.142, (2.093)

2.165 (2.096)

0.925-0.950 2.093, (2.076)

2.099, (2.099)

2.101, (2.095)

2.105, (2.102)

1.18-1.22

(1.19-1.20)

2.086(5),

2.098(5),

2.094(5),

2.096(5)

U-Oendo-U 107.5 (107.1) 107.1 (107.5) 106.6

U-Ocis-U 106.6 (106.6) 106.6 (106.8) 106.4

U-Nimine 2.539-2.563

(2.538-2.540)

0.370 2.537-2.562

(2.543-2.572)

0.38-0.41

(0.37-0.40)

2.490-2.515

U-Npyrrole 2.432-2.480

(2.468-2.471)

0.435 2.457-2.474

(2.443-2.484)

0.52-0.55

(0.51-0.57)

2.420-2.442

Page 302: Relativistic Quantum Chemistry Applied to Actinides

280

O-Si 1.688 (1.688) 0.829 1.703 (1.700) 1.04 (1.04) 1.663

Figure 8.5: Molecular orbitals of primary σ- and π- character in the unrestricted singlet state of

1a: (a) α-(HOMO-27) with energy of –0.333 a.u. and contributions of 27% endo-oxo 2p, 13 %

exo-oxo 2p, 3 % cis-oxo 2p and 13 % U-5f; (b) β-(HOMO-27), with energy of –0.333 a.u. and

contributions of 25 % endo-oxo 2p, 11 % exo-oxo 2p, 3 % cis-oxo 2p and 13% U-5f. These σ-

type orbitals extend across the U2O2 core; (c) α-HOMO-28 with energy of –0.334 a.u. and

Page 303: Relativistic Quantum Chemistry Applied to Actinides

281

contributions of 34 % cis-oxo 2p and 9 % endo-oxo 2p, 5% U-5f and 6 % U-6d; (d) β-HOMO-

28, with energy of –0.334 a.u. and contributions of 37 % cis-oxo 2p, 9 % endo-oxo 2p, 5 % U-5f

and 5 % U-6d. These orbitals depict the weaker π-type interaction across the U2O2 core.

Figure 8.6: The (a) HOMO-172 (bonding with respect to the two U atoms) with an energy of -

1.094 a.u. and (b) its antibonding counterpart, LUMO+35, both containing contributions from U-

5f and O-2s orbitals.

theory.45 There is a need to further explore the presence of any covalent interaction between the

actinide centers in this and similar complexes. The calculated Mayer bond orders for the U-U

internuclear distances are intriguing at best.

Page 304: Relativistic Quantum Chemistry Applied to Actinides

282

Formation of 1

The formation of 1 can occur with a uranyl-Pacman, UO2(py)(H2L), starting complex. It is

however unclear whether the insertion of the second uranyl group of 1 proceeds via a hexavalent

or pentavalent complex. There is also the possibility of formation via the oxo-silylated complex,

Me3SiOUO(py)(H2L) as dictated by the oxo-functionalized nature of 1. This is interesting as

UO2(py)(H2L) and its pentavalent and oxo-functionalized pentavalent derivatives most likely

provide different reactivities for the transamination of the second amine site of UO2(py)(H2L).

The incoming uranyl group is inserted into the Pacman complexes in a cis- orientation in order to

achieve the butterfly-shaped U2O4 core found in 1. We have used the uranyl silylamide complex,

UO2[N(SiMe3)2]2(py)2, to supply the incoming uranyl units as it is the same reagent that was

used in the experimental synthesis of 1. These reactions are depicted in Scheme 8.1 and their

calculated energies are presented in Table 8.2. The B3LYP/RECP and PBE/AE/4-component

calculations are in agreement that the insertion of a cis-uranyl entity into the hexavalent complex

is an endothermic reaction while it is exothermic, and significantly so, for the pentavalent

complex. This is not surprising given the greater Lewis basicity of the oxo atoms in the

pentavalent state. For the oxo-silylated complex, Reaction 3, we find that the reaction energy is

exothermic but less so than the case of the pentavalent complex, Reaction 2, Table 8.2. The

formation of a binuclear complex from a pentavalent oxo-silylated complex is expected to be

aided by increased Lewis basicity of the endo- oxo atoms of the Pacman complexes in

comparison to the hexavalent complex, Table 8.3. Examination of the calculated atomic charges

on the hexavalent, pentavalent and oxo-silylated starting complexes reveals an increase in the

polarization of the U-O bonds in the oxo-functionalized complexes, indicative of an increase in

ionicity and a decrease in bond covalency. This effect is especially more pronounced in the

Page 305: Relativistic Quantum Chemistry Applied to Actinides

283

functionalized U-O bonds. The increased acidity of the uranium atoms in the functionalized

complexes and to a smaller extent the greater basicity of the endo- oxo atoms suggest increased

reactivity with Lewis bases and acids respectively, Table 8.2. It should be noted that, although

the changes in the calculated atomic charges are minute, they are in line with the previous work

of Wang et al, who studied NUN and NUNH (proton-functionalization of a uranyl analog).46

Overall, the reaction energies for these transamination reactions suggest that reductive oxo-

silylation has a lesser effect on the basicity of the endo- oxo atom than reduction to the

pentavalent uranyl complex. We can basically conclude from the results presented in Table 8.2

that the formation of the U2O4 core of 1 most likely involves the pentavalent uranyl Pacman

complex.

Scheme 8.1: Insertion of a cis-type UO2 group into the empty, un-transaminated site of the

uranyl Pacman complex. X is R3Si, R3C and H for the oxo- silylated, alkylated and protonated

complexes respectively.

Page 306: Relativistic Quantum Chemistry Applied to Actinides

284

Table 8.2: Calculated reaction free energies (ΔG298, kcal/mol)47

required to transaminate the

unoccupied amine site of several Pacman complexes with a cis-uranyl group. These reactions

result in the formation of butterfly U2O4 motif similar to that found in complex 1. R is SiMe3.

PBE/L1 B3LYP/RECP

Reactions Gas Gas Pya

1. UO2(py)(H2L) + UO2(NR2)2(py)2 → (UO2)2L + 3py + 2HNR2 4.8 13.5 9.3

2. [UO2(py)(H2L)]- + UO2(NR2)2(py)2 → [(UO2)2L]

- + 3py +

2HNR2 (Pentavalent)

-12.8 -12.5 -14.2

3. Me3SiOUO(py)(H2L) + UO2(NR2)2(py)2 → (Me3SiOUO2UO)L +

3py + 2HNR2 (Oxo-functionalized pentavalent)

-3.1 -3.8 -5.1

a In pyridine solvent using gas phase geometries and the PCM/UA0 approach

Table 8.3: Calculated atomic charges on the uranium and oxo- atoms of the uranyl Pacman

complex, UO2(py)(H2L), and its pentavalent and reductively oxo-functionalized derivatives

obtained at the PBE/AE/4-component level (and at the B3LYP/RECP level).

U(VI)O2(py)(H2L) [U(V)O2(py)(H2L)]− HOU(V)O(py)(H2L) Me3SiOU(V)O(py)(H2L)

U 1.38 (1.76) 1.24 (1.52) 1.42 (1.76) 1.61 (1.87)

Oendo -0.39 (-0.62) -0.46 (-0.74) -0.42 (-0.67) -0.42 (-0.67)

Oexo -0.27 (-0.46) -0.32 (-0.56) -0.39 (-0.66) -0.50 (-0.78)

Page 307: Relativistic Quantum Chemistry Applied to Actinides

285

Electronic structure analysis of the lithiated monomer complexes, 2M and 3M

The structural properties of the lithium-functionalized monomer complexes of the yttrium

and samarium dimer salts were optimized at the PBE/AE/4-component level. The calculated

structural properties are compiled in Table 8.4. The structure of the Y monomer complex is

shown in Figure 8.4. For the Sm monomer complex, the quintet (4fα55fβ

1) and septet (4fα

55fα

1)

electronic states were considered. The septet state was found to be slightly more stable than the

quintet indicating that ferromagnetic coupling of the Sm and U centers is slightly favored. The

energy difference between these states, 0.6 kcal/mol, is quite small. However, the structural

parameters for the quintet state are in much better agreement with the experimental crystal

structure, Table 8.4. The UO2 stretching vibrations in the calculated IR spectrum of the yttrium

complex at 698 and 757 cm-1

are in good agreement with the peaks found at 764 (UO2

asymmetric stretch) and 725 (UO2 asymmetric stretch) cm-1

in the measured FTIR spectrum.

In the Sm and Y monomer complexes, the calculated Mayer bond orders indicate

significant double bond characters in the U-O1 and U-O2 bonds, Table 8.4. On the other hand, the

M-Cl, M-O1 and U-Cl interactions are single bonds with appreciable ionic characters. The bond

orders for the U-O bonds in these complexes are much lower than was obtained for the

hexavalent uranyl Pacman complex, UO2(py)(H2L).30

The loss of a whole bond order is due to

the interaction of the oxo atoms with the Li and lanthanide atoms, as the triple bond character of

the U-O bonds in the uranyl group is preserved on single electron reduction of UO2(py)(H2L) to

UO2(py)(H2L)-. This phenomenon was also observed in Wang et al.‟s study of NUN and

NUNH.46

Examination of the molecular orbitals of the yttrium complex however reveals the σ

and π orbital manifolds similar to those of uranyl groups. The symmetry (Lewis acid-base

interaction between the oxo- and electropositive metal atoms on both sides of the uranium center

Page 308: Relativistic Quantum Chemistry Applied to Actinides

286

Table 8.4: The structural properties (bond lengths in Å and angles in degrees) for the monomer

complexes obtained at the PBE/L1 level. The calculated Mayer bond orders are given in

parenthesis. The experimental data are taken from Reference 31.

Sm-monomer

Calc.

Expt.

Y-monomer

Calc. Expt.

Septet Quintet

Li-O2 1.91 (0.53) 1.89 (0.58) 1.90 (0.57)

U-O1 1.86 (2.02) 1.90 (1.83) 1.91 1.92 (1.79)

U-O2 1.87 (1.79) 1.89 (1.75) 1.86 1.89 (1.76)

M-O1 2.45 (0.34) 2.32 (0.51) 2.24 (0.53)

M-Cl 2.94 (0.29) 2.87 (0.35) 2.80 (0.36)

U-Cl 2.74 (0.78) 2.79 (0.67) 2.79 (0.66)

U-M 3.70 3.65 3.63 3.63

Li-O2-U2 175.0 175.0 173.4 175.0

O1-U1-O2 176.4 176.2 174.7 175.2

U-O1-M 118.1 120.1 122.3

U-Cl-M 82.0 81.1 80.7

Page 309: Relativistic Quantum Chemistry Applied to Actinides

287

) preserves the optimal geometry for interaction of the valence orbitals of the uranium and

oxygen atoms. However depletions of the Y-O1 2p atomic contributions to the π orbitals are

characteristic of oxo-functionalized uranyl systems and explain the weakening of the U-O1

bonds.

Electronic structure analysis of the dimer complexes, 2 and 3

We also performed geometry optimizations on 2 and 3, the direct representations of the

Sm and Y dimer complexes produced experimentally.31

For the yttrium dimer complex, 2, the

triplet and broken symmetry singlet states were calculated as being essentially iso-energetic

(energy difference of 0.1 kcal/mol) in the gas phase at the PBE/L1 level. The optimized structure

of the broken symmetry singlet state obtained at the PBE/L1 level is shown in Figure 8.3. The

experimental IR spectrum for this complex has dominant peaks at 897, 836, 791, 751, 722, 702,

668, 621, 603 and 569 cm-1

. These are well matched by the calculated vibrational modes at 570,

600, 605, 623, 685, 691, 709, 730, 737, 792, 851 and 884 cm-1

, in order. The peaks between 570

and 709 cm-1

in the calculated spectrum corresponds to several motions of the U and O atoms in

the U2O4 core, amongst which are the U-O1 stretching and flexing of the U2O2 core.

For the samarium dimer complex, 3, the 4fα55fα

15fβ

14fβ

5 singlet state (i.e.

antiferromagnetic coupling between the two uranium f1 centers and ferromagnetic interaction

between the Sm and U centers, 3s) was calculated to be 9.2 kcal/mol more stable than the

4fβ55fα

15fα

14fβ

5 state (total multiplicity of 9, 3n). The greater stability of the antiferromagnetic

singlet state is in contrast to the experimental magnetic results which reveal greater

superexchange interaction between the Sm and U centers rather than between the two uranium

centers.

Page 310: Relativistic Quantum Chemistry Applied to Actinides

288

The calculated and experimental U-M distances in the dimer complexes, Table 8.5, are

significantly larger than those in the Li(py)3(μ-Cl) complexes, Table 8.4. This is most likely due

to the presence of the bridging chloride groups in the monomers which occupy an acute angle

above the U-M distance, despite similar M-O distances in the monomers and dimers. Like in the

Table 8.5: The calculated and experimental structural parameters (bond lengths in Å and angles

in degrees) for the dimer complexes obtained at the PBE/L1 level. The Mayer bond orders are

given in parenthesis.

Sm-Dimer, 3 Y-Dimer, 2

Singlet, 3s Nonet, 3n Expt. Singlet Expt.

U-O1 1.88 (2.00) 1.92 (1.83) 1.89 1.94 (1.79) 1.92

U-O2 1.95 (1.71) 1.98 (1.63) 1.94 1.98 (1.61) 1.97

U-O2′ 2.34 (0.63) 2.31 (0.71) 2.35 2.30 (0.73) 2.32

M-O1 2.32 (0.35) 2.25 (0.53) 2.24 2.18 (0.55) 2.16

U-U 3.47 3.47 (0.34) 3.47 3.47 3.45

U-M 4.20 4.18 4.12 4.11 4.07

O1-U1-O2 175.4 175.4 174.4 175.6 175.3

U1-O1-M 176.1 178.6 174.5 176.2 177.3

Page 311: Relativistic Quantum Chemistry Applied to Actinides

289

monomer complexes, the U-O1 and U-O2 distances in the dimers have double bond characters.

The CCIs across the U2O2 core, depicted by the U-O2′ bonds, are however weaker and

more ionic bonds than the other U-O bonds. The calculated U-U distances in the dimer

complexes are about 3.47 Å. This is shorter than the sum of the covalent radii of two singly

bonded uranium atoms, 3.92 Å but slightly longer than the value of 3.37 Å recently obtained by

us in the butterfly oxo-silylated complex in which a U2O2 core is formed from two pentavalent

uranyl units within the same pyrrolic macrocycle as used here, Table 8.1.28

Conclusions

In this work we have studied the structural and electronic properties of three binuclear

U2O4 complexes with U(V)/U(V) configuration using scalar relativistic DFT calculations. The

ground electronic states of these complexes as well as the possibilities for U-U interactions were

also examined. For the formation of the butterfly-shaped complex, 1, the Lewis basicities of the

endo- oxo atoms of the oxo-functionalized complexes are intermediate between those of the

hexavalent and pentavalent uranyl complexes. As a result of this, the transamination reactions

leading to the butterfly-shaped U2O4 motif are more exothermic for the pentavalent and oxo-

functionalized pentavalent species than the hexavalent complexes. The reaction energies point to

the possible involvement of a pentavalent intermediate in the synthesis of 1.

The f1 uranium centers in 1 were found to be anti-ferromagnetically coupled in agreement

with low temperature magnetic experiments. In contrast, for 2 and 3, antiferromagnetic coupling

amongst the U centers and ferromagnetic coupling between the actinide and lanthanide centers is

predicted by the PBE/L1 all-electron basis set approach in disagreement with the experimental

work. However, this is only severe in the Sm-dimer complex, 3, with the ferromagnetic and

Page 312: Relativistic Quantum Chemistry Applied to Actinides

290

antiferromagnetic U-U or M-U coupling being degenerate (< 1 kcal/mol) for the yttrium dimer,

2. It should however be noted that we were simply unable to converge to the 4fβ55fα

15fβ

14fα

5

electronic state of the Sm-dimer that was indicated by the experimental magnetic measurements.

The bonding characteristics reveal that the U-O bonds in 2 and 3 as well as their monomeric

precursors, are mainly of double bond character, representing a loss of a full bond order, in

comparison to UO2(py)(H2L), upon oxo-metallation by either Li (for the monomeric precursors,

2M and 3M) or lanthanide metals (for 2 and 3).

References

1. Arnold, P. L.; Love, J. B.; Patel, D., Coordin. Chem. Rev. 2009, 253 (15-16), 1973-1978.

2. Ephritikhine, M., Dalton T. 2006, (21), 2501-2516.

3. Fortier, S.; Hayton, T. W., Coordin. Chem. Rev. 2010, 254 (3-4), 197-214.

4. Graves, C. R.; Kiplinger, J. L., Chem. Commun. 2009, (26), 3831-3853.

5. Melfi, P. J.; Kim, S. K.; Lee, J. T.; Bolze, F.; Seidel, D.; Lynch, V. M.; Veauthier, J. M.;

Gaunt, A. J.; Neu, M. P.; Ou, Z.; Kadish, K. M.; Fukuzumi, S.; Ohkubo, K.; Sessler, J. L., Inorg.

Chem. 2007, 46 (13), 5143-5145.

6. Nocton, G.; Horeglad, P.; Vetere, V.; Pecaut, J.; Dubois, L.; Maldivi, P.; Edelstein, N.

M.; Mazzanti, M., J. Am. Chem. Soc. 2010, 132 (2), 495-508.

7. Seidel, D.; Lynch, V.; Sessler, J. L., Angew. Chem. Int. Edit. 2002, 41 (8), 1422-1425.

8. Sessler, J. L.; Gorden, A. E. V.; Seidel, D.; Hannah, S.; Lynch, V.; Gordon, P. L.;

Donohoe, R. J.; Tait, C. D.; Keogh, D. W., Inorg. Chim. Acta 2002, 341, 54-70.

9. Sessler, J. L.; Melfi, P. J.; Pantos, G. D., Coordin. Chem. Rev. 2006, 250 (7-8), 816-843.

10. Sessler, J. L.; Seidel, D.; Vivian, A. E.; Lynch, V.; Scott, B. L.; Keogh, D. W., Andw.

Chem. Int. Edit. 2001, 40 (3), 591-594.

Page 313: Relativistic Quantum Chemistry Applied to Actinides

291

11. Sessler, J. L.; Vivian, A. E.; Seidel, D.; Burrell, A. K.; Hoehner, M.; Mody, T. D.;

Gebauer, A.; Weghorn, S. J.; Lynch, V., Coordin. Chem. Rev. 2001, 216, 411-434.

12. Ikeda, A.; Hennig, C.; Tsushima, S.; Takao, K.; Ikeda, Y.; Scheinost, A. C.; Bernhard,

G., Inorg. Chem. 2007, 46 (10), 4212-4219.

13. Mizuoka, K.; Grenthe, I.; Ikeda, Y., Inorg. Chem. 2005, 44 (13), 4472-4474.

14. Mizuoka, K.; Ikeda, Y., Inorg. Chem. 2003, 42 (11), 3396-3398.

15. Mizuoka, K.; Ikeda, Y., Radiochim. Acta 2004, 92 (9-11), 631-635.

16. Mizuoka, K.; Kim, S. Y.; Hasegawa, M.; Hoshi, T.; Uchiyama, G.; Ikeda, Y., Inorg.

Chem. 2003, 42 (4), 1031-1038.

17. Berthet, J. C.; Nierlich, M.; Ephritikhine, M., Angew. Chem. Int. Edit. 2003, 42 (17),

1952-1954.

18. Arnold, P. L.; Patel, D.; Blake, A. J.; Wilson, C.; Love, J. B., J. Am. Chem. Soc. 2006,

128 (30), 9610-9611.

19. Arnold, P. L.; Patel, D.; Wilson, C.; Love, J. B., Nature 2008, 451 (7176), 315-U3.

20. Arnold, P. L.; Pecharman, A. F.; Love, J. B., Angew. Chem. Int. Edit. 2011, 50 (40),

9456-9458.

21. Schnaars, D. D.; Wu, G.; Hayton, T. W., J. Am. Chem. Soc. 2009, 131 (48), 17532-

17533.

22. Hayton, T. W.; Wu, G., Inorg. Chem. 2009, 48 (7), 3065-3072.

23. Burdet, F.; Pecaut, J.; Mazzanti, M., J. Am. Chem. Soc. 2006, 128 (51), 16512-16513.

24. Mougel, V.; Horeglad, P.; Nocton, G.; Pecaut, J.; Mazzanti, M., Angew. Chem. Int. Edit.

2009, 48 (45), 8477-8480.

Page 314: Relativistic Quantum Chemistry Applied to Actinides

292

25. Mougel, V.; Horeglad, P.; Nocton, G.; Pecaut, J.; Mazzanti, M., Chem-Eur. J. 2010, 16

(48), 14365-14377.

26. Natrajan, L.; Burdet, F.; Pecaut, J.; Mazzanti, M., J. Am. Chem. Soc. 2006, 128 (22),

7152-7153.

27. Nocton, G.; Horeglad, P.; Pecaut, J.; Mazzanti, M., J. Am. Chem. Soc. 2008, 130 (49),

16633-16645.

28. Arnold, P. L.; Jones, G. M.; Odoh, S. O.; Schreckenbach, G.; Magnani, N.; Love, J. B.,

Nature Chem. 2012, 4 (3), 221-227.

29. Arnold, P. L.; Blake, A. J.; Wilson, C.; Love, J. B., Inorg. Chem. 2004, 43 (26), 8206-

8208.

30. Berard, J. J.; Schreckenbach, G.; Arnold, P. L.; Patel, D.; Love, J. B., Inorg. Chem. 2008,

47 (24), 11583-11592.

31. Arnold, P. L.; Hollis, E.; White, F. J.; Magnani, N.; Caciuffo, R.; Love, J. B., Angew.

Chem. Int. Edit. 2011, 50 (4), 887-890.

32. Arnold, P. L.; Holis, E.; Love, J. B.; Magnani, N.; Colineau, E.; Caciuffo, R.; Edelstein,

N.; Maron, L.; Castro, L.; Yahia, A.; O., O. S.; G., S., 2012.

33. Odoh, S. O.; Schreckenbach, G., J. Phys. Chem. A 2011, 115 (48), 14110–14119.

34. Becke, A. D., J. Chem. Phys. 1993, 98 (7), 5648-5652.

35. Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J., J. Phys. Chem. 1994, 98

(45), 11623-11627.

36. Miertus, S.; Scrocco, E.; Tomasi, J., Chem. Phys. 1981, 55 (1), 117-129.

37. Miertus, S.; Tomasi, J., Chem. Phys. 1982, 65 (2), 239-245.

38. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74 (6), 1245-1263.

Page 315: Relativistic Quantum Chemistry Applied to Actinides

293

39. Küchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100 (10), 7535-7542.

40. Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman,

J. R.; Scalmani, G.; Barone, V.; Mennucci, B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li,

X.; Hratchian, H. P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.; Ehara,

M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima, T.; Honda, Y.; Kitao, O.;

Nakai, H.; Vreven, T.; Montgomery, J., J. A.; Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J.

J.; Brothers, E.; Kudin, K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;

Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, M.; Millam, N. J.; Klene,

M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.; Jaramillo, J.; Gomperts, R. E.; Stratmann,

O.; Yazyev, A. J.; Austin, R.; Cammi, C.; Pomelli, J. W.; Ochterski, R.; Martin, R. L.;

Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich, S.;

Daniels, A. D.; Farkas, O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,

Revision A.2, Gaussian, Inc.: Wallingford CT, 2009.

41. Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; F., W. NBO Version 3.1.

42. Bridgeman, A. J.; Cavigliasso, G.; Ireland, L. R.; Rothery, J., J. Chem. Soc. Dalton 2001,

(14), 2095-2108.

43. Laikov, D. N.; Ustynyuk, Y. A., Russ. Chem. B+ 2005, 54 (3), 820-826.

44. Pan, Q.-J.; Shamov, G. A.; Schreckenbach, G., Chem-Eur. J. 2010, 16 (7), 2282-2290.

45. Gagliardi, L.; Roos, B. O., Nature 2005, 433 (7028), 848-851.

46. Wang, X. F.; Andrews, L.; Vlaisavljevich, B.; Gagliardi, L., Inorg. Chem. 2011, 50 (8),

3826-3831.

47. Shamov, G. A.; Schreckenbach, G., J. Phys. Chem. A 2005, 109 (48), 10961-10974.

Page 316: Relativistic Quantum Chemistry Applied to Actinides

294

Preface to Chapter 9

This chapter is based on manuscript that would soon be submitted for publication in the journal

“Inorganic Chemistry”. The full citation of the paper is as follows:

Samuel O. Odoh, Bert W. de Jong and Georg Schreckenbach, “Theoretical Study of a

Gas-Phase Binuclear Uranyl Hydroxo Complex, (UO2)2(OH)5-.” Inorganic Chemistry,

2012, to be submitted.

The work in this chapter provides a detailed description of our search for the lowest energy

structures of the bis-uranyl complex, (UO2)2(OH)5-. This complex was identified in laser ablation

studies of uranium trioxide solids. The relative energies of the various structures were calculated

out at the density functional theory (DFT) level and using ab initio wavefunction approaches.

All the DFT as well as the full-MP2 calculations in this chapter were carried out by Samuel O.

Odoh. The frozen-core MP2 and CCSD(T) calculations were carried out by Bert de Jong. The

manuscript was prepared together with the other authors.

Copyright permissions have been obtained from all the other authors. Bert de Jong is affiliated

with the Pacific Northwest National Laboratory.

Page 317: Relativistic Quantum Chemistry Applied to Actinides

295

Chapter 9: Theoretical Study of a Gas-Phase Binuclear Uranyl

Hydroxo Complex, (UO2)2(OH)5-.

Abstract

The low energy structures and bonding of gas-phase (UO2)2(OH)5-, a product of laser

ablation studies on uranium trioxide solids, have been studied using density functional theory

(DFT) and ab initio methods while employing scalar relativistic effective core potentials. As

(UO2)2(OH)5- is a member of the (UO2)2(OH)n

4-n, series of complexes, the structures of the di, tri,

tetra and hexa-hydroxo bis-uranyl complexes were also determined with particular emphasis on

structures featuring cation-cation interactions between the uranyl groups. We found that the most

stable structures obtained for (UO2)2(OH)5- feature this type of interactions. This was confirmed

at the MP2 and CCSD(T)//MP2 levels. Analysis of the bonding in the structures of (UO2)2(OH)5-

shows that the UO2(OH)2 and UO2(OH)3- units are held together mainly by electrostatic effects.

In addition, there is no evidence for covalent interactions between the actinide atoms even in

structures where the U-U distances are significantly lesser than the sum of the covalent radii of

the actinide centers. The calculated IR vibrational frequencies provide signature probes that can

be used in differentiating the low-energy structures and in experimentally confirming the

existence of the structures featuring cation-cation interactions.

Introduction

There is a synergistic convergence between the application of computational approaches

to the study of actinide complexes1-24

and the resurgence of synthetic actinide chemistry. Novel

actinide complexes produced via “wet” chemical synthesis are being reported regularly. A few

Page 318: Relativistic Quantum Chemistry Applied to Actinides

296

examples of these are stable U(V) complexes25

, imido-analogs of actinyl species26-27

among

others. In addition to these wet chemical approaches, there is a long tradition of identifying

gaseous actinide compounds produced from laser-ablated solids by mass spectrometric

analyses.28-52

The transfer of actinide compounds from the solvent phase into the gaseous phase

has also been demonstrated. In addition, the species produced after the laser ablation and

solution-to-gas transfer processes have been reacted with other compounds (such as nebulized

alcohols, water and oxygen).33-36, 39, 47-49

In several cases, the products of such reactions were

found to themselves be new actinide species. As an example, the discovery of gas-phase

hypercoordinated complexes of the actinyl dications has been reported by van Stipdonk et al.47

A

collision-induced dissociation (CID) approach has also been applied to gaseous actinide

systems.34, 48

The fragmentation patterns and product yields have provided insights into the redox

chemistries of actinyl complexes as well as the nature of the chemical bonds in these compounds.

Marcalo et al. identified multinuclear uranates with molecular formulas ranging between

UOn- and U14On

- in their ablative work on solid uranium trioxide.

53 Anionic species, such as

U2O7H- and U3O10H

-, containing hydrogen atoms were also detected in their mass spectrometric

data. They noted that the hydrogen atoms were produced from water or hydroxyl groups in their

samples. More recently, a comprehensive mass-spectrometric study on the gaseous species

produced from laser ablation of solid uranium trioxide in the presence of water has lead to the

identification of highly hydroxylated (UO2)n(OH)2n+1- and (UO2)n(OH)2n(O2)

- clusters.

54 The

fragmentation patterns of these clusters were also examined using state of the art CID

techniques. A recent ablation study of titanium dioxide solids resulted in the detection of similar

oxy-hydroxide anions, (TiO2)x(H2O)yOH- and (TiO2)x(H2O)y(O2)

-..55

Page 319: Relativistic Quantum Chemistry Applied to Actinides

297

One of the gaseous species identified in the mass spectrometric work is (UO2)2(OH)5-. It

is an interesting complex as it is one of the simplest of the bis-uranyl hydroxo complexes

identified, in addition to having the largest yields in the mass spectra of the ablated gases.54

In

many cases, the experimental identification and characterization of gaseous actinide compounds

have been accompanied or followed by detailed theoretical examination of their structural and

electronic properties as well as their fragmentation or reaction pathways.6, 40, 44, 50

The relatively

small size of (UO2)2(OH)5- means it is relatively tractable in terms of computational expense. It

should however be noted that given its bis-uranyl nature (and the multi-nuclear nature of the

(UO2)n(OH)2n+1- and (UO2)n(OH)2nO2

- clusters in general) as well as the relatively large number

of ligands, there is a need to examine a large number of low-energy structures while searching

for the global minimum point on the potential energy surface.

The crystal structures of UO356

and UO3.H2O57

are known to possess cation-cation

interactions (CCIs) between neighboring uranyl groups. CCIs are intermolecular interactions

between cationic species. In the case of bis-actinyl species, these are usually formed through oxo

atoms. CCIs are found in UO3 as the uranium atoms in the crystal structure have only oxo-type

neighbors. As the (UO2)n(OH)2n+1- anionic clusters are formed from UO3 (and reaction of the

ablated gases with water or hydroxyl), it is important that structural frameworks retaining the

CCIs found in the solid structure are considered in any search for the lowest energy structures for

these complexes. Another perspective involves examining the individual components of the

(UO2)2(OH)5- complex. The UO2

2+ group contains a d

0f0 uranium atom, is rigorously linear and

is known to possess very inert oxo atoms.58-60

On the other hand, the hydroxo ligand is a strong σ

and π donor. The coordination of electron donors to the equatorial region of the uranyl moiety

generally results in an increase in the Lewis basicities of the uranyl oxo atoms. This increased

Page 320: Relativistic Quantum Chemistry Applied to Actinides

298

Lewis basicities of the oxo atoms has been found to favor the formation of CCIs between uranyl

groups or with other cations.61

From this point of view, the role of CCIs in the (UO2)n(OH)2n+1-

complexes would be expected to increase as the number of hydroxo ligands are increased. It can

then be concluded that a search for the lowest energy structures of (UO2)2(OH)5- is incomplete

without a search for structures containing CCIs between the uranyl groups.

On the other hand, the structural and electronic properties of bis-uranyl hydroxo-aquo

complexes, formed at high pH values in aqueous solutions, have been examined with

experimental and theoretical approaches. In their calculations, Tsushima et al. obtained a μ2-

dihydroxo structure for (UO2)2(H2O)6(OH)22+

complex as well as a μ-hydroxo structure for the

(UO2)2(H2O)8(OH)3+

complex in aqueous solution.19

The μ2-dihydroxo structure that was

obtained for (UO2)2(H2O)6(OH)22+

is in good agreement with the crystallographic work of

Aberg.62

In this structure, the uranyl groups are bridged by hydroxo ligands and their equatorial

coordination sphere is satisfied by aquo ligands. Similar μ-hydroxo bridged structures have also

been obtained for other actinide complexes.63-65

It should however be noted that there is no

guarantee that these μ2-dihydroxo structures would be the most stable motifs for the highly

hydroxylated complexes belonging to the (UO2)n(OH)2n+1- series in the gaseous phase.

We report here a theoretical study of the structural features and electronic properties of

the low energy structures of (UO2)2(OH)5-. Particular emphasis was given to the search of stable

structures featuring CCIs between the uranyl groups in this complex. The structure of other

members of the (UO2)2(OH)n4-n

series of complexes were also examined to further determine if

CCIs are pervasive in gaseous bis-uranyl hydroxide complexes. The density functional theory

(DFT), second order Møller-Plesset perturbation (MP2) and coupled cluster approaches were

used in this work.

Page 321: Relativistic Quantum Chemistry Applied to Actinides

299

Computational Details

All the calculations were performed with the NWChem 6.0 package of programs.66

Scalar

relativistic calculations were carried out with the Stuttgart small-core effective core potential for

the uranium atom.67-69

The valence basis associated with this pseudopotential is of 10s9p5d5f

contraction while all-electron DFT optimized valence triple-ζ polarized (TZVP) basis sets were

used for the oxygen and hydrogen atoms.70

This combination of basis set and pseudopotentials

have been previously used in calculating accurate structural parameters, vibrational frequencies

and reaction energies of actinide and transition metal complexes.5, 7, 71-72

The geometries of many

possible structural motifs of (UO2)2(OH)5- and other members of the (UO2)2(OH)n

4-n series of

complexes were optimized using the hybrid B3LYP functional. These structural motifs were

obtained by rigorous examination of many possible structures without symmetry constraints.

This approach allows for a sampling of the potential energy surface with the aim of detecting the

lowest energy structures. Structures containing CCIs between the uranyl groups and those

featuring bridging hydroxo- ligands were included in this search. In addition, structures

containing equatorial aquo ligands (while maintaining the molecular formula of U2O9H5-) were

considered. To further ascertain the relative energies of the lowest energy structures, their

geometries were also optimized with the local density approximation (LDA), the half and half

functional of Becke, as implemented in NWChem (BeckeHandH), and MP2. The U-U distances

between the two uranium atoms of the title complex and several of the structures studied in this

work are less than 3.92 Å (twice the covalent radius of a uranium atom73

). In addition some

ligands bonded to each uranium atom are within the second coordination sphere of the adjacent

uranium atom. These two effects suggest the possibility of substantial long-range correlation

effects on the structure and relative energies of these complexes. For this reason, single point

Page 322: Relativistic Quantum Chemistry Applied to Actinides

300

calculations on the MP2 optimized geometries were carried out with the coupled cluster,

CCSD(T), approach as well as a variety of modified density functionals: CAM-B3LYP74

,

BeckeHandH as well as the B3LYP functional corrected for dispersion effects using Grimme‟s

third scheme, B3LYP-D3.75

To characterize the bonding in the different structures, the Mayer-

Mulliken bond order76

and Mulliken atomic charges were calculated at the B3LYP/TZVP level.

Results and Discussion

Structures of (UO2)2(OH)5-. The three lowest energy structures obtained for (UO2)2(OH)5

- at the

B3LYP/TZVP level are presented in Figure 9.1. The calculated relative energies of these

structures obtained while employing a variety of density functional and ab initio methods are

collected in Table 9.1. At the B3LYP/TZVP level, the μ2-dihydroxo structure was found to be

the least stable of these three structures. The two lower-energy structures that were calculated

both feature CCIs between their uranyl groups and are essentially degenerate. One of these CCI

structures possesses a bridging hydroxo group between the uranium centers, Figure 9.1 in

addition to an oxo linkage. This structure is labeled as μ-hydroxo-CCI, Figure 9.1. The uranyl

group whose oxo atom participates in the CCI essentially lies in the equatorial plane of the other

uranyl group. In this structure, each uranyl group possesses two equatorially coordinated pendant

hydroxo ligands in addition to the bridging hydroxo group. The bonds between the bridging

hydroxo group and the uranium atoms were calculated as 2.28 and 2.51 Å, Table 9.2. As would

be expected, the bridging hydroxo ligand is much more strongly bonded to the equatorial uranyl

group, whose oxo atom is involved in the CCI, Figure 9.1. The U-Oyl bonds for the three oxo

atoms not involved in the CCIs were calculated to be about 1.80 Å long, Table 9.2. This is within

the range calculated for UO2(OH)2, 1.79 Å and UO2(OH)3-, 1.81 Å, at the B3LYP/TZVP level,

Table 9.3. A similar case was also observed for the bonds between the uranium atoms and the

Page 323: Relativistic Quantum Chemistry Applied to Actinides

301

pendant hydroxo ligands, the U-OHpendant bonds, Table 9.2. The U-OCCI bond for the equatorial

uranyl group and U1-OCCI between the uranyl groups were calculated as 1.84 and 2.62 Å

respectively.

Figure 9.1: The three low energy structures of (UO2)2(OH)5- obtained at the B3LYP/TZVP

level.

The second CCI structure is labeled as μ-hydroxo-di-CCI, Figure 9.1. The uranyl groups in

this structure are not fully parallel to each other with their orientations such that the pendant

hydroxo groups are all in the equatorial region. However, the fifth hydroxo group is a bridging

hydroxo group coordinated to both uranyl groups. The U-OCCI bonds for both uranyl groups

(bonds between the uranium atoms and the oxo groups involved in CCIs) are about 1.86-1.87 Å

long, Table 9.2. These are much longer than the free U-Oyl bonds which were calculated to be

1.80-1.81 Å in length. As noted, this was also the case for the μ-hydroxo-CCI structure, Table

9.2, suggesting involvement in CCIs results in elongation of U-Oyl bonds by 0.04-0.07 Å. The

U1-OCCI bonds between the uranyl groups were calculated as 2.56 and 2.63 Å, confirming the

presence of some asymmetry in this structure, Table 9.2. Attempts at removing this asymmetry

Page 324: Relativistic Quantum Chemistry Applied to Actinides

302

Table 9.1: Calculated relative energies (kcal/mol) of the three low energy structures of

(UO2)2(OH)5-.

μ2-dihydroxo μ-hydroxo-CCI μ-hydroxo-di-CCI

LDA 8.92 9.10 0.00

B3LYP 1.83 0.00 0.17

MP2//B3LYP 4.29 3.89 0.00

CAM-B3LYP 1.12 0.00 0.71

BeckeHandH 12.97 7.82 0.00

MP2 6.31 5.45 0.00

B3LYP//MP2 5.87 2.75 0.00

B3LYP-D3//MP2 4.58 2.54 0.00

CAM-B3LYP//MP2 6.49 3.69 0.00

BeckeHandH//MP2 9.02 4.67 0.00

CCSD(T)//MP2 5.13 3.04 0.00

by stipulating a C2 axis through the μ-hydroxo ligand increases the total energy by 1.42 kcal/mol

and results in a transition state structure at the B3LYP/TZVP level. The pendant U-OH bonds in

the UO2(OH)2 groups of this structure were calculated to be 2.16 -2.17 Å long. They are within

the range of the U-OH bond lengths calculated for UO2(OH)2 (2.11 Å) and UO2(OH)3- (2.22 Å),

similar to the case for the μ-hydroxo-CCI.

The third low-energy framework obtained at the B3LYP/TZVP level appears to formed

through the agglomeration of UO2(OH)2 and UO2(OH)3- via two bridging hydroxo ligands. This

structure is labeled as the μ2-dihydroxo structure, Figure 9.1. The U-Oyl bond lengths in this

Page 325: Relativistic Quantum Chemistry Applied to Actinides

303

Table 9.2: Calculated bond lengths (Å) and bond orders of the low energy structures of

(UO2)2(OH)5- obtained at the B3LYP/TZVP level.

μ2-dihydroxo μ-hydroxo-CCI μ-hydroxo-di-CCI

Length Order Length Order Length Order

U-Oyl 1.79-1.81 2.02-2.08 1.80 2.04 1.80-1.81 2.04

U-OCCI 1.84 1.71 1.86-1.87 1.63-1.68

U1-OCCI 2.62 0.25 2.56

2.63

0.32

0.29

U-OHbridging 2.26-2.29

2.56

0.56-0.58 2.28

2.51

0.49

0.21

2.44

2.36

0.37

0.46

U-OHpendant 2.18-2.19 0.27 2.17-2.19 0.87-0.91 2.16-2.17 0.88-0.91

U-U 3.84 0.02 3.78 0.04 3.39 0.10

structure were calculated to be about 1.79-1.81 Å. The bonds between the bridging hydroxo

uranyl groups are significantly longer for the uranyl group with two pendant hydroxo groups.

The U-OH bonds for the pendant hydroxo groups were predicted to be about 2.18-2.19 Å long,

Table 9.2. This μ2-dihydroxo structure was calculated to be 1.66 and 1.82 kcal/mol respectively

Page 326: Relativistic Quantum Chemistry Applied to Actinides

304

higher in energy than the μ-hydroxo-di-CCI and μ-hydroxo-CCI structures at the B3LYP/TZVP

level, Table 9.1.

It is interesting that two of the lowest energy structures of (UO2)2(OH)5- obtained at the

B3LYP/TZVP level feature CCIs between the uranyl groups. CCIs have been observed in the

solid-state structures of UO356

and UO3.H2O56

. The U-Oyl distances in solid state UO3.H2O range

from 2.38 Å to 2.56 Å.56

The U1-OCCI distances in μ-hydroxo-CCI and μ-hydroxo-di-CCI

structures were however calculated to be between 2.56 and 2.63 Å, suggesting some retention of

the general structure found in the solid phase after ablation and reaction in the gas phase. The

longer U1-OCCI distances in the gas-phase structures are most likely due to the highly

hydroxylated nature of the (UO2)2(OH)5- complex.

The calculated U-U distances in these structures were all found to be less than the sum of

the covalent radii of the uranium centers (3.92 Å).73

The U-U distances were calculated as 3.39,

3.78 and 3.84 Å for the μ-hydroxo-di-CCI, μ-hydroxo-CCI and μ2-dihydroxo structures

respectively, Table 9.2. The contraction of the U-U distance as the degree of CCIs is increased is

consistent with the geometrical arrangements of the different structures.

Relative Energies of the Most Stable Structures. The calculated energy differences between

these stable structures are found to be rather small at the B3LYP/TZVP level, Table 9.1. There is

a possibility of substantial long-range effects on the calculated relative-energies of these

structures due to the short U-U distances as well as the fact that some ligands bonded to each

uranium atom are within the second-coordination sphere of the other uranium atom. There is also

a need to confirm the stability of the structures featuring CCIs by employing other theoretical

approaches. Towards this end, structural optimizations were carried out with the LDA,

Page 327: Relativistic Quantum Chemistry Applied to Actinides

305

BeckeHandH and CAM-B3LYP functionals as well as at the MP2 level. In addition, single point

calculations at the MP2 optimized geometries were carried out using a coupled-cluster,

CCSD(T), approach and several of long-range corrected density functionals, Table 9.1.

Optimization at the MP2/TZVP level breaks the degeneracy found between the μ-

hydroxo-di-CCI and μ-hydroxo-CCI structures at the B3LYP/TZVP level, Table 9.1. Similarly,

the LDA, BeckeHandH and CAM-B3LYP functionals predict the μ-hydroxo-di-CCI structure to

be lower in energy than the μ-hydroxo-CCI structure. This is confirmed by the single-point

energy calculations with the CCSD(T) approach at the MP2 optimized geometries. These

calculations confirm the rather high-energy nature of the μ2-dihydroxo structure (4.5-9.0

kcal/mol) in comparison to the two structures featuring CCIs. Methodologically, the general

trend in the relative energies obtained at the CCSD(T)//MP2 and MP2 levels are replicated by

the BeckeHandH functional. The half and half functional however predicts the μ2-dihydroxo

structure to be significantly higher in energy than is obtained with the ab initio approaches, Table

9.1. Optimization with the local density approximation (LDA) correctly predicts the μ-hydroxo-

di-CCI structure as the lowest energy structure, although the μ-hydroxo-CCI and μ2-dihydroxo

structures were found to be iso-energetic. Lastly, optimization with CAM-B3LYP, yields

disappointing results for the overall trend as well as the magnitude of the relative energies when

compared to the MP2 and CCSD(T)//MP2 results, Table 9.1.

Calculated Vibrational Frequencies. The calculated harmonic infra-red (IR) frequencies

associated with the uranyl and hydroxo groups in the three most stable structures of

(UO2)2(OH)5- are presented in Table 9.3. Firstly, the O-H stretching of the hydroxo groups are

found at around 3791 cm-1

in all the three structures. This is typical of uranium complexes

containing hydroxo groups and is also present in the mono-nuclear uranyl hydroxo complexes.77

Page 328: Relativistic Quantum Chemistry Applied to Actinides

306

The U-OH stretching modes were calculated to be around 546-573 and 597-612 cm-1

in

UO2(OH)2 and UO2(OH)3- complexes respectively at the B3LYP/TZVP level. The calculated

values for UO2(OH)2 are in good agreement with the experimental (547-569 cm-1

) and

theoretical (534-572 cm-1

) values obtained by Wang et al.77

In (UO2)2(OH)5-, the U-O-H

stretching modes of the pendant hydroxo ligands were calculated to be between 560 and 580 cm-

1 for the μ2-dihydroxo structure. These modes are found at 513-555 cm

-1 for the μ-hydroxo-di-

CCI and 533-577 cm-1

for the μ-hydroxo-CCI structures, Figure 9.2. Overall the peaks centered

at around 570 cm-1

are of modest intensities in the μ2-dihydroxo and μ-hydroxo-CCI structures.

The intense peaks between 420 and 520 cm-1

correspond to the U-O-H bending and H-O-U-O-H

asymmetric stretching modes among others. The U-O-H vibrational modes of the bridging

hydroxo groups are however found at much higher wavenumbers. These were predicted to be at

748 and 734 cm-1

in the μ2-dihydroxo structure at the B3LYP/TZVP level, Table 9.3. They are

found at around 630 and 649 cm-1

in the μ-hydroxo-CCI and μ-hydroxo-di-CCI structures

respectively.

The uranyl groups in mononuclear uranyl complexes are generally involved in a set of IR

active asymmetric and an IR inactive symmetric stretching vibrations. The symmetric uranyl

stretching modes were calculated to be at 883 and 832 cm-1

for UO2(OH)2 and UO2(OH)3-

respectively at the B3LYP/TZVP level while the asymmetric stretching modes were respectively

calculated at about 955 and 905 cm-1

, Table 9.3. Wang et al. reported values of 920 and 944 cm-1

from their experimental and theoretical work on UO2(OH)2 respectively.77

The calculated IR

spectra for the three low energy structures of (UO2)2(OH)5-, Figure 9.2, show the presence of

Page 329: Relativistic Quantum Chemistry Applied to Actinides

307

Table 9.3: Calculated IR vibrational frequencies (cm-1

) of the low energy structures of

(UO2)2(OH)5- obtained at the B3LYP/TZVP level. The normalized intensities are given in

parenthesis.

μ2-dihydroxo μ-hydroxo-CCI μ-hydroxo-di-CCI

Uranyl Modes

Asymmetric*

O1a-U1-O1b 921 (0.9) 899 (0.6) 871 (0.2)

O2a-U1-O2b 895 (0.2) 882 (0.7) 867 (1.0)

Symmetric*

O1a-U1-O1b 843 (0.0) 828 (0.0) 783 (0.2)

O2a-U1-O2b 823 (0.0) 783 (0.2) 732 (0.1)

Bridging Hydroxo 748 (0.2), 734 (0.1) 631 (0.1) 649 (0.1)

Pendant Hydroxo 587 (0.0), 578 (0.2),

573 (0.1), 563 (0.2)

577 (0.1), 571 (0.2),

537 (0.0), 534 (0.0)

555 (0.0), 535 (0.1),

532 (0.0)

Mixed hydroxo 518 (0.5) 520 (0.1) 524 (0.0)

* The O1a-U1-O1b and O2a-U2-O2b stretching modes correspond to stretching of each of the uranyl

groups.

Page 330: Relativistic Quantum Chemistry Applied to Actinides

308

Figure 9.2: Calculated IR spectra for the μ2-dihydroxo (black), μ-hydroxo-CCI (blue) and μ-

hydroxo-di-CCI (red) structures of (UO2)2(OH)5- obtained using the B3LYP functional.

four uranyl stretching modes between 750-950 cm-1

. These modes can be divided into two sets

between 750-860 and 860-950 cm-1

. The sets found at higher wavenumbers involve asymmetric

uranyl stretching while the sets at lower wavenumbers correspond to symmetric uranyl stretching

modes. It should however be noted that there is some variability in the degree of coupling

between the stretching modes of the individual uranyl groups. Interestingly, the lowest uranyl

stretching mode (783 cm-1

) in the μ-hydroxo-CCI structure and the two lowest (783 and 732 cm-

1) in the μ-hydroxo-di-CCI structures were predicted to yield peaks of modest IR intensities.

These peaks correspond largely to the stretching vibrations of the U-OCCI bonds. The modest

Page 331: Relativistic Quantum Chemistry Applied to Actinides

309

intensities of these vibrational modes are in contrast to the weak IR activity of the symmetric

stretching vibrations of the uranyl group in the μ2-dihydroxo structure, Figure 9.2. This provides

a tantalizing suggestion of the possibility of confirming the presence of CCIs in the ground state

structure of (UO2)2(OH)5-. A measurement of the IR and Raman spectra of gaseous (UO2)2(OH)5

-

could provide equivocal confirmation of the existence or otherwise of CCIs in the gas phase

structure, as the symmetric stretching modes of the uranyl group in the μ2-dihydroxo structure

are significantly weaker (IR) and are found at about 50 cm-1

higher wavenumbers than those in

the μ-hydroxo-CCI and μ-hydroxo-di-CCI structures. In addition, the peaks corresponding to

wagging of the bridging hydroxo ligands at around 750 cm-1

for the μ2-dihydroxo structure and

those corresponding to the stretching of the U-Oyl bonds in the CCI structures peaks (near 800

cm-1

) can also be used as probes to distinguish the μ2-dihydroxo structure from those featuring

CCIs. Overall, the calculated relative intensities of the peaks obtained with the LDA and

BeckeHandH functionals confirm the usefulness of these vibrational modes.

Bonding. The Mayer-Mulliken bond order has been very useful in the description of bonding in

inorganic molecules.76, 78

The U-Oyl bonds in the bare uranyl group were calculated to have a

bond order of 2.42, Table 9.4, indicating significant triple bond character. Addition of hydroxo

ligands to the mononuclear uranyl species gradually decreases the multiplicity of the U-Oyl

bonds to 2.01 in UO2(OH)3-. A similar elongation and accompanying decrease in the bond orders

of the U-OH bonds is seen with larger number of equatorial hydroxo ligands.

Examination of the calculated bond orders in the structures of (UO2)2(OH)5- allows for

several conclusions, Table 9.2. Firstly, for the pendant hydroxo ligands, the U-OH bond orders

are between 0.86 and 0.92, intermediate between the values obtained for UO2(OH)2, 1.01, and

UO2(OH)3-, 0.81 respectively, Table 9.4. The calculated bond orders of the U-Oyl bonds in the

Page 332: Relativistic Quantum Chemistry Applied to Actinides

310

Table 9.4: Calculated bond lengths (Å), bond orders and vibrational frequencies (cm-1

) of a few

gas-phase mono-nuclear uranium complexes obtained at the B3LYP/TZVP level.

Bond Lengths (Å) Bond Orders Frequencies (cm-1

)

U-Oyl U-O U-OH U-Oyl U-O U-OH

UO22+

1.701 2.42 1049, 1149

UO2(OH)+ 1.745 2.010 2.27 1.21 955, 1033

UO2(OH)2 1.785 2.114 2.12 1.01 883, 955

UO2(OH)3- 1.811 2.221 2.01 0.81 832, 905

UO3 1.807 1.849 2.13 2.18 789, 893, 893

uranyl groups of the μ2-dihydroxo structure and the free U-Oyl bonds of the CCI structures range

between 2.02 and 2.08. In contrast for the oxo atoms involved in CCIs, the U-OCCI bond orders

are significantly decreased, ranging in value between 1.63 and 1.71. The reduction in bond order

is similar but lesser than the significant decrease of up to a full bond order usually observed upon

reductive oxo-functionalization.50, 79

Regarding the oxo and hydroxo bridges across the uranium

centers, the U1-OCCI bonds possess bond orders between 0.25 and 0.32 while the U-OHbridging

bonds have bond orders between 0.21 and 0.58.

Lastly, the short U-U distances, especially in the μ-hydroxo-di-CCI structure, are most

likely due to the hard-hard nature of the uranium-oxygen interactions. We draw this conclusion

Page 333: Relativistic Quantum Chemistry Applied to Actinides

311

because the calculated bond orders across the U-U distances are negligible (0.02-0.10). This

certainly suggests that covalent interactions between the actinide centers are essentially non-

existent.

The dissociation of the three low-energy structures of (UO2)2(OH)5- into UO2(OH)2 and

UO2(OH)3- were calculated to be endothermic at the B3LYP/TZVP level. The energy required

ranges from 50.0 kcal/mol in the case of the μ2-dihydroxo structure to 51.6 kcal/mol for the μ-

hydroxo-di-CCI structure. This confirms the stability of the structures, with the bridging U-OH

and U-O′CCI bonds sufficiently holding the two units together. Given the low bond orders

obtained for the bridging U-OH and U-O′CCI bonds, we examined the calculated Mulliken

charges of the involved atoms. The effective charges on the oxygen atoms of the pendant

hydroxo groups are generally between -0.72 and -0.75 while the charges for the free uranyl oxo

atoms are between -0.51 and -0.56. In contrast, the uranyl oxo atoms involved in CCIs possess

effective charges between -0.60 and -0.63 while for the bridging hydroxo groups the effective

charges were calculated to be between -0.78 and -0.85. It appears that the bridging oxo and

hydroxo groups are overall more polarized than the free oxo and pendant hydroxo groups

respectively. A reflection of this is seen in the fact that the uranium centers with oxo atoms

participating in CCIs possess larger effective charges. This is in line with lower overlap with the

oxo atoms (smaller bond order for the U-OCCI bonds in contrast to the free U-Oyl bonds) as well

as larger electrostatic interaction of the CCI oxo atoms with the adjacent uranium atoms (very

low bond orders for the U1-OCCI bonds).

Other Structures of (UO2)2(OH)5-. There are many other possible structures for (UO2)2(OH)5

-.

We present a few of them in Figure 9.3. The complexes with aquo ligands, A, B and C, are of

significantly higher energies than the hydroxylated ones. This is the case at both the B3LYP and

Page 334: Relativistic Quantum Chemistry Applied to Actinides

312

MP2//B3LYP level. Single point MP2 calculations on the B3LYP/TZVP optimized geometries

appear sufficient as they also predict the removal of the degeneracy between the μ-hydroxo-CCI

and μ-hydroxo-di-CCI that was obtained at the B3LYP level, Table 9.1. The aquo complexes are

essentially (UO3)n(OH)-.nH2O adducts formed as intermediates to the hydroxylated complexes.

Other structures include those with only one bridging oxo, D, or one hydroxo group, E, which

are higher in energy than the structures in Figure 9.1, a reflection of the fact that the hydroxo

bridges as well as the CCIs confer stability on the structural framework. The relative energy of E

with respect to the μ-hydroxo-di-CCI structure increases to 12.3 kcal/mol at the MP2//B3LYP

level. This is unsurprising and is in line with the results presented in Table 9.1, MP2 favors the

CCI structures. Of the remaining structures presented in Figure 9.3, G is particularly interesting.

It is very similar to the μ-hydroxo-CCI structure, with a pendant hydroxo group now converted

into a second bridging ligand. This structure was predicted to be about 13.7 and 2.4 kcal/mol

higher in energy than the μ-hydroxo-di-CCI structure at the B3LYP and MP2//B3LYP levels

respectively. The implication of this is that it is even more stable than the μ-hydroxo-CCI

structure at the MP2//B3LYP level. Attempts at re-optimizing the structure of G at the MP2 level

however results in it collapsing to the μ-hydroxo-di-CCI structure.

Structures of other (UO2)2(OH)n4-n

complexes. (UO2)2(OH)5- is a member of the

(UO2)2(OH)n4-n

series of complexes and was identified in a mass spectrometric study.54

To

determine the role of CCIs in the gas phase structures of the other complexes in this series, the

structures of (UO2)2(OH)22+

, (UO2)2(OH)3+, (UO2)2(OH)4, and (UO2)2(OH)6

2- were all optimized

using similar approaches and basis sets employed for (UO2)2(OH)5-. The most stable structure for

(UO2)2(OH)22+

was found to be a μ2-dihydroxo structure at the B3LYP/TZVP level. A μ-

hydroxo-CCI structure, Figure 9.4, was calculated to be about 4.2 kcal/mol higher in energy than

Page 335: Relativistic Quantum Chemistry Applied to Actinides

313

the μ2-dihydroxo structure at this level, Table 9.5. In addition, another structure in which two

UO2(OH)+ groups interact via their oxo atoms was predicted to be 16.5 kcal/mol higher in energy

than the μ2-dihydroxo structure. This structure is labeled as di- CCI. The general trend in the

relative stabilities of the μ2-dihydroxo, μ-dihydroxo-CCI and di-CCI structures of (UO2)2(OH)22+

obtained with the B3LYP functional is in agreement with those obtained at the MP2 and

CCSD(T)//MP2 levels, Table 9.5.

Figure 9.3: Other possible structure of (UO2)2(OH)5-. Their energies at the B3LYP (and

MP2//B3LYP) levels are given relative to that of the μ-hydroxo-di-CCI structure.

Page 336: Relativistic Quantum Chemistry Applied to Actinides

314

The μ2-dihydroxo framework was also found as the lowest energy structures of the tetrahydroxo,

(UO2)2(OH)4, and hexahydroxo, (UO2)2(OH)62-

, complexes at the B3LYP/TZVP level, Figure

9.4. Structurally, the trend towards increasingly longer U-Oyl and U-OH bonds with increasing

numbers of equatorial hydroxo ligands is evident in these complexes. Energetically, the μ-

hydroxo-CCI structures were calculated to be higher in energy than the μ2-dihydroxo ones by

about 4.05 and 6.11 kcal/mol for the tetra- and hexahydroxo complexes at the B3LYP/TZVP

level respectively, Table 9.5. Similar to the case in the dihydroxo complex, the di-CCI structures

were found at much higher energies (11.13 and 18.37 kcal/mol for the tetra- and hexahydroxo

complexes). The trends in the relative energies of these three structures calculated at the MP2

and CCSD(T)//MP2 levels are in good agreement with those obtained at the B3LYP level. We

note that the tetrahydroxo complexes were predicted to be stable with respect to fragmentation

into two units of UO2(OH)2. The decompositions of the di-CCI, μ-hydroxo-CCI and μ2-

dihydroxo structures of (UO2)2(OH)4 were calculated to be endothermic by about 33.19, 40.00

and 43.12 kcal/mol respectively at the B3LYP/TZVP level. The stability of these structures is

understandable given that the charge on each monomeric UO2(OH)2, unit is zero. This also

indicates that while the di-CCI structure of the tetrahydroxo complex, Figure 9.4, is significantly

higher in energy than the lowest energy structure, the CCIs can still bind the monomer units

together with a significant amount of energy (33.19 kcal/mol). We also explored a variety of

other structures for (UO2)2(OH)4 and (UO2)2(OH)62-

, amongst which are those with a butterfly

shaped U2O4 core, similar to the recently reported complex,80

and those similar to structure G in

Figure 9.3.

Page 337: Relativistic Quantum Chemistry Applied to Actinides

315

Table 9.5: Relative energies (kcal/mol) of the low energy structures of (UO2)2(OH)n4-n

, (n=2, 3,

4 and 6) obtained at the DFT and ab initio levels.

B3LYP MP2 CCSD(T)//MP2

(UO2)2(OH)22+

μ2-dihydroxo 0.00 0.00 0.00

μ-hydroxo-CCI 4.15 3.89 7.60

di-CCI 16.51 12.66 35.86

(UO2)2(OH)3+

μ2-dihydroxo 0.00 0.48 0.00

μ-hydroxo-CCI 0.80 1.36 1.19

μ-hydroxo-di-CCI 3.49 0.00 3.55

(UO2)2(OH)4

μ2-dihydroxo 0.00 0.00 0.00

μ-hydroxo-CCI 4.04 3.76 4.03

di-CCI 11.13 9.68 10.98

(UO2)2(OH)62-

μ2-dihydroxo 0.00 0.00 0.00

Page 338: Relativistic Quantum Chemistry Applied to Actinides

316

μ-hydroxo-CCI 6.11 7.63 8.70

di-CCI 18.37 17.97 20.09

Figure 9.4: Low energy structures of the bis-uranyl hydroxo complexes, (UO2)2(OH)n4-n

.

TETRAHYDROXO

TRIHYDROXO

HEXAHYDROXO

PENTAHYDROXO

DIHYDROXO

Page 339: Relativistic Quantum Chemistry Applied to Actinides

317

The trihydroxo complex is similar to the title complex, (UO2)2(OH)5-, in that they both

have an odd number of hydroxo groups. This, like in the pentahydroxo complex, allows for a

bridging hydroxide in between two UO2(OH)+ entities. For this complex, Figure 9.4, the μ2-

dihydroxo structure was calculated to be only 0.80 kcal/mol higher in energy than the μ-hydroxo-

CCI structure at the B3LYP/TZVP level while the μ-hydroxo-di-CCI structure was calculated to

be 3.50 kcal/mol higher in energy, Table 9.5. The small relative energies of these structures are

in distinct contrast to the relative energies of similar structures in the di-, tetra- and hexahydroxo

species, Table 9.5. Overall this suggests that the increased stability of structural frameworks

featuring CCIs between the uranyl groups begins in the trihydroxo complex. It would appear that

the odd number of hydroxo ligands confers increased stabilization by a bridging ligand across the

uranyl groups. The μ-hydroxo-di-CCI and μ2-dihydroxo structures were found to be the most

stable structure at the MP2 and CCSD(T)//MP2 levels respectively. The μ-hydroxo-CCI and μ-

hydroxo-di-CCI structures were calculated to be about 1.19 and 3.55 kcal/mol higher energy than

the μ2-dihydroxo structure at the CCSD(T)//MP2 level. The smaller energy differences between

these structures and the greater stabilities of the CCI structures in the pentahydroxo complex,

Table 9.1, indicate the role of molecular symmetry (with respect to the pendant and bridging

ligands).

Conclusions

We have examined the structural and electronic properties of several structures of the gas

phase bis-uranyl hydroxo complex, [(UO2)2(OH)5]-, using a range of DFT and ab initio methods.

A similar approach was also employed in the search for the stable structures of the

(UO2)2(OH)n4-n

(n=2, 3, 4 and 6) complexes. Particular emphasis was given to low energy

structures featuring CCIs between the two uranyl groups.

Page 340: Relativistic Quantum Chemistry Applied to Actinides

318

Structurally, the bis-uranyl complexes are similar to their uranyl counterparts. The

progressive elongation of the U-Oyl and U-OH bonds as the number of equatorial hydroxo

ligands in the [(UO2)2(OH)n]4-n

complexes is similar to the observation in the uranyl complexes.

A search of the low energy structures for the [(UO2)2(OH)n]4-n

complexes shows that the μ2-

dihydroxo structures are significantly more stable than those featuring CCIs for the dihydroxo,

tetrahydroxo and hexahydroxo complexes. In contrast, structures featuring uranyl CCIs were

however found to be more stable than the μ2-dihydroxo structures for the trihydroxo and

pentahydroxo complexes. A combination of the role of hydroxo ligands as σ-donors increasing

the Lewis basicity of the uranyl oxo atoms and symmetry is most likely responsible for the

greater stability of the CCI structures in the gaseous trihydroxo and pentahydroxo complexes. It

remains to be seen whether the CCI structural arrangements have been overlooked in aqueous

phase chemistry.

For the calculated structural parameters of (UO2)2(OH)5-, the U-O bonds for the oxo-

atoms involved in CCIs are significantly elongated in comparison to their free counterparts. This

is similar to the popular case of oxo-functionalized pentavalent and hexavalent uranyl

complexes. The formation of CCIs creates discrepancies in the calculated IR vibrational

frequencies associated with the stretching of the U-Oyl bonds suggesting that the existence or

otherwise of the CCI structures can be experimentally confirmed by measuring the IR spectrum

of gas-phase (UO2)2(OH)5-. In addition, signature U-O-H stretching of the bridging hydroxo

group at around 750 cm-1

and U-Oyl stretching peaks around 800 cm-1

should allow for a

conclusive differentiation of the CCI and dihydroxo structures.

Page 341: Relativistic Quantum Chemistry Applied to Actinides

319

Examination of the bonding in [(UO2)2(OH)5]- reveals the ionic nature of the μ-hydroxo

and CCIs oxo bridges. Negligible covalent interactions were found between the two actinide

centers even in structures with significantly short U-U distances.

References

1. Barros, N.; Maynau, D.; Maron, L.; Eisenstein, O.; Zi, G. F.; Andersen, R. A.,

Organometallics 2007, 26, 5059.

2. Bolvin, H.; Wahlgren, U.; Moll, H.; Reich, T.; Geipel, G.; Fanghanel, T.; Grenthe, I., J.

Phys. Chem. A 2001, 105, 11441.

3. Brynda, M.; Wesolowski, T. A.; Wojciechowski, K., J. Phys. Chem. A 2004, 108, 5091.

4. Clavaguera-Sarrio, C.; Vallet, V.; Maynau, D.; Marsden, C. J., J. Chem. Phys. 2004, 121,

5312.

5. de Jong, W. A.; Harrison, R. J.; Nichols, J. A.; Dixon, D. A., Theor. Chem. Acc. 2001,

107, 22.

6. Gagliardi, L.; Roos, B. O., Chem. Phys. Lett. 2000, 331, 229.

7. Groenewold, G. S.; Gianotto, A. K.; McIlwain, M. E.; Van Stipdonk, M. J.; Kullman, M.;

Moore, D. T.; Polfer, N.; Oomens, J.; Infante, I.; Visscher, L.; Siboulet, B.; De Jong, W. A., J.

Phys. Chem. A 2008, 112, 508.

8. Groenewold, G. S.; Oomens, J.; de Jong, W. A.; Gresham, G. L.; McIlwain, M. E.; Van

Stipdonk, M. J., Phys. Chem. Chem. Phys. 2008, 10, 1192.

9. Gutowski, K. E.; Cocalia, V. A.; Griffin, S. T.; Bridges, N. J.; Dixon, D. A.; Rogers, R.

D., J. Am. Chem. Soc. 2007, 129, 526.

Page 342: Relativistic Quantum Chemistry Applied to Actinides

320

10. Hemmingsen, L.; Amara, P.; Ansoborlo, E.; Field, M. J., J. Phys. Chem. A 2000, 104,

4095.

11. Kozimor, S. A.; Yang, P.; Batista, E. R.; Boland, K. S.; Burns, C. J.; Clark, D. L.;

Conradson, S. D.; Martin, R. L.; Wilkerson, M. P.; Wolfsberg, L. E., J. Am. Chem. Soc. 2009,

131, 12125.

12. Nocton, G.; Horeglad, P.; Vetere, V.; Pecaut, J.; Dubois, L.; Maldivi, P.; Edelstein, N.

M.; Mazzanti, M., J. Am. Chem. Soc. 2010, 132, 495.

13. Real, F.; Vallet, V.; Marian, C.; Wahlgren, U., J. Chem. Phys. 2007, 127.

14. Schreckenbach, G.; Hay, P. J.; Martin, R. L., Inorg. Chem. 1998, 37, 4442.

15. Schreckenbach, G.; Hay, P. J.; Martin, R. L., J. Comput. Chem. 1999, 20, 70.

16. Schreckenbach, G.; Shamov, G. A., Accounts Chem. Res. 2010, 43, 19.

17. Shamov, G. A.; Schreckenbach, G., J. Am. Chem. Soc. 2008, 130, 13735.

18. Sundararajan, M.; Campbell, A. J.; Hillier, I. H., J. Phys. Chem. A 2008, 112, 4451.

19. Tsushima, S.; Rossberg, A.; Ikeda, A.; Muller, K.; Scheinost, A. C., Inorg. Chem. 2007,

46, 10819.

20. Vallet, V.; Wahlgren, U.; Schimmelpfennig, B.; Moll, H.; Szabo, Z.; Grenthe, I., Inorg.

Chem. 2001, 40, 3516.

21. Vazquez, J.; Bo, C.; Poblet, J. M.; de Pablo, J.; Bruno, J., Inorg. Chem. 2003, 42, 6136.

22. Wahlin, P.; Danilo, C.; Vallet, V.; Real, F.; Flament, J. P.; Wahlgren, U., J. Chem.

Theory Comput. 2008, 4, 569.

23. Wander, M. C. F.; Kerisit, S.; Rosso, K. M.; Schoonen, M. A. A., J. Phys. Chem. A 2006,

110, 9691.

24. Yahia, A.; Arnold, P. L.; Love, J. B.; Maron, L., Chem. Commun. 2009, 2402.

Page 343: Relativistic Quantum Chemistry Applied to Actinides

321

25. Graves, C. R.; Kiplinger, J. L., Chem. Commun. 2009, 3831.

26. Hayton, T. W.; Boncella, J. M.; Scott, B. L.; Batista, E. R.; Hay, P. J., J. Am. Chem. Soc.

2006, 128, 10549.

27. Hayton, T. W.; Boncella, J. M.; Scott, B. L.; Palmer, P. D.; Batista, E. R.; Hay, P. J.,

Science 2005, 310, 1941.

28. Andrews, L.; Liang, B. Y.; Li, J.; Bursten, B. E., Angew. Chem. Int. Edit. 2000, 39,

4565.

29. Chertihin, G. V.; Andrews, L.; Neurock, M., J. Phys. Chem-US 1996, 100, 14609.

30. Chien, W.; Anbalagan, V.; Zandler, M.; Van Stipdonk, M.; Hanna, D.; Gresham, G.;

Groenewold, G., J. Am. Soc. Mass. Spectr. 2004, 15, 777.

31. Cornehl, H. H.; Wesendrup, R.; Diefenbach, M.; Schwarz, H., Chem-Eur. J. 1997, 3,

1083.

32. Di Santo, E.; Santos, M.; Michelini, M. C.; Marcalo, J.; Russo, N.; Gibson, J. K., J. Am.

Chem. Soc. 2011, 133, 1955.

33. Gibson, J. K., Int. J. Mass Spectrom. 2002, 214, 1.

34. Gibson, J. K.; Marcalo, J., Coordin. Chem. Rev. 2006, 250, 776.

35. Gresham, G. L.; Gianotto, A. K.; Harrington, P. D.; Cao, L. B.; Scott, J. R.; Olson, J. E.;

Appelhans, A. D.; Van Stipdonk, M. J.; Groenewold, G. S., J. Phys. Chem. A 2003, 107, 8530.

36. Hunt, R. D.; Andrews, L., J. Chem. Phys. 1993, 98, 3690.

37. Hunt, R. D.; Yustein, J. T.; Andrews, L., J. Chem. Phys. 1993, 98, 6070.

38. Jackson, G. P.; Gibson, J. K.; Duckworth, D. C., J. Phys. Chem. A 2004, 108, 1042.

39. Jackson, G. P.; King, F. L.; Goeringer, D. E.; Duckworth, D. C., J. Phys. Chem. A 2002,

106, 7788.

Page 344: Relativistic Quantum Chemistry Applied to Actinides

322

40. Lyon, J. T.; Andrews, L.; Malmqvist, P. A.; Roos, B. O.; Yang, T. X.; Bursten, B. E.,

Inorg. Chem. 2007, 46, 4917.

41. Marcalo, J.; Gibson, J. K., J. Phys. Chem. A 2009, 113, 12599.

42. Michelini, M. D.; Marcalo, J.; Russo, N.; Gibson, J. K., Inorg. Chem. 2010, 49, 3836.

43. Michelini, M. D.; Russo, N.; Sicilia, E., J. Am. Chem. Soc. 2007, 129, 4229.

44. Pereira, C. C. L.; Marsden, C. J.; Marcalo, J.; Gibson, J. K., Phys. Chem. Chem. Phys.

2011, 13, 12940.

45. Ricks, A. M.; Gagliardi, L.; Duncan, M. A., J. Phys. Chem. Lett. 2011, 2, 1662.

46. Santos, M.; Marcalo, J.; Leal, J. P.; de Matos, A. P.; Gibson, J. K.; Haire, R. G., Int. J.

Mass Spectrom. 2003, 228, 457.

47. Van Stipdonk, M. J.; Chien, W.; Anbalagan, V.; Bulleigh, K.; Hanna, D.; Groenewold, G.

S., J. Phys. Chem. A 2004, 108, 10448.

48. Van Stipdonk, M. J.; Chien, W.; Anbalagan, V.; Gresham, G. L.; Groenewold, G. S., Int.

J. Mass Spectrom. 2004, 237, 175.

49. Van Stipdonk, M. J.; Chien, W.; Bulleigh, K.; Wu, Q.; Groenewold, G. S., J. Phys.

Chem. A 2006, 110, 959.

50. Wang, X. F.; Andrews, L.; Vlaisavljevich, B.; Gagliardi, L., Inorg. Chem. 2011, 50,

3826.

51. Zhou, M. F.; Andrews, L.; Ismail, N.; Marsden, C., J. Phys. Chem. A 2000, 104, 5495.

52. Zhou, M. F.; Andrews, L.; Li, J.; Bursten, B. E., J. Am. Chem. Soc. 1999, 121, 9712.

53. Marcalo, J.; Santos, M.; de Matos, A. P.; Gibson, J. K., Inorg. Chem. 2009, 48, 5055.

54. Groenewold, G.; de Jong, W. A., Personal Communication 2011.

55. Barthen, N.; Millon, E.; Aubriet, F., J. Am. Soc. Mass. Spectr. 2011, 22, 508.

Page 345: Relativistic Quantum Chemistry Applied to Actinides

323

56. Siegel, S.; Hoekstra, H.; Sherry, E., Acta Crystallogr. 1966, 20, 292.

57. Siegel, S.; Tani, B.; Viste, A.; Hoekstra, H. R., Acta Crystallogr. B-Stru. 1972, B 28,

117.

58. Denning, R. G., Struct. Bond. 1992, 79, 215.

59. Denning, R. G., J. Phys. Chem. A 2007, 111, 4125.

60. Denning, R. G.; Snellgrove, T. R.; Woodwark, D. R., Mol. Phys. 1979, 37, 1109.

61. Fortier, S.; Hayton, T. W., Coordin. Chem. Rev. 2010, 254, 197.

62. Aberg, M., Acta. Chem. Scand. A 1978, 32, 101.

63. Henry, N.; Lagrenee, M.; Loiseau, T.; Clavier, N.; Dacheux, N.; Abraham, F., Inorg.

Chem. Commun. 2011, 14, 429.

64. Thuery, P., Acta Crystallogr. C 2007, 63, M54.

65. Thuery, P.; Masci, B., Acta Crystallogr. C 2002, 58, m556.

66. Valiev, M.; Bylaska, E. J.; Govind, N.; Kowalski, K.; Straatsma, T. P.; Van Dam, H. J. J.;

Wang, D.; Nieplocha, J.; Apra, E.; Windus, T. L.; de Jong, W., Comput. Phys. Commun. 2010,

181, 1477.

67. Bergner, A.; Dolg, M.; Kuchle, W.; Stoll, H.; Preuss, H., Mol. Phys. 1993, 80, 1431.

68. Kuchle, W.; Dolg, M.; Stoll, H.; Preuss, H., Mol. Phys. 1991, 74, 1245.

69. Kuchle, W.; Dolg, M.; Stoll, H.; Preuss, H., J. Chem. Phys. 1994, 100, 7535.

70. Godbout, N.; Salahub, D. R.; Andzelm, J.; Wimmer, E., Can. J. Chem.-Rev. Can. Chim.

1992, 70, 560.

71. Aubriet, F.; Gaumet, J. J.; de Jong, W. A.; Groenewold, G. S.; Gianotto, A. K.; McIlwain,

M. E.; Van Stipdonk, M. J.; Leavitt, C. M., J. Phys. Chem. A 2009, 113, 6239.

Page 346: Relativistic Quantum Chemistry Applied to Actinides

324

72. Groenewold, G. S.; Gianotto, A. K.; Cossel, K. C.; Van Stipdonk, M. J.; Oomens, J.;

Polfer, N.; Moore, D. T.; de Jong, W. A.; McIlwain, M. E., Phys. Chem. Chem. Phys. 2007, 9,

596.

73. Cordero, B.; Gomez, V.; Platero-Prats, A. E.; Reves, M.; Echeverria, J.; Cremades, E.;

Barragan, F.; Alvarez, S., Dalton T. 2008, 2832.

74. Yanai, T.; Tew, D. P.; Handy, N. C., Chem. Phys. Lett. 2004, 393, 51.

75. Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H., J. Chem. Phys. 2010, 132.

76. Bridgeman, A. J.; Cavigliasso, G.; Ireland, L. R.; Rothery, J., J. Chem. Soc. Dalton 2001,

2095.

77. Wang, X. F.; Andrews, L.; Li, J., Inorg. Chem. 2006, 45, 4157.

78. Bridgeman, A. J.; Cavigliasso, G., Faraday Discuss. 2003, 124, 239.

79. Berard, J. J.; Schreckenbach, G.; Arnold, P. L.; Patel, D.; Love, J. B., Inorg. Chem. 2008,

47, 11583.

80. Arnold, P. L.; Jones, G. M.; Odoh, S. O.; Schreckenbach, G.; Magnani, N.; Love, J. B.,

Nature Chem. 2011, 4, 221

Page 347: Relativistic Quantum Chemistry Applied to Actinides

325

Chapter 10: Summary and Future Studies of Actinide Complexes

10.1 Summary

The studies presented in this thesis have provided a great amount of information

regarding the use and performance of computational techniques for studying actinide complexes.

The performance of modern DFT and relativistic ECPs were benchmarked with an aim of

determining the degrees of accuracy that could be expected when they are used for calculations

on actinide systems. These approaches were then used in studying the speciation and properties

of actinyl hydroxides in solution (Chapter 4), monomeric uranium peroxides in solution (Chapter

5), uranium fluorides in the gas and aqueous phases as well as in a cell-wall protein (Chapter 7),

reduction of U(V) complexes on geochemical surfaces such as rutile and anatase (Chapter 6),

novel and highly stable pentavalent uranium complexes formed by reductive oxo-

functionalization of axial oxo atoms in a bimetallic bis-uranium dioxo motif (Chapter 8) and

lastly the structure and properties of a novel homobimetallic U(VI) hydroxide complexes in the

gas-phase (Chapter 9).

The linkages between these works revolve around the role and existence of actinide

species in the environment. Based on the studies compiled in this thesis, we can make a general

statement like this: Some components of highly radioactive nuclear waste decompose by

emission of α-particles. This induces the α-hydrolysis of water giving hydrogen peroxide. It is

important to study monomeric uranyl peroxide complexes, Chapter 5, as they are the building

blocks of crystalline polynuclear species characterized by Burns et al.1-7

They might be of great

importance at low H2O2 concentrations, confirmed by the existence of only thin peroxide films on

depleted uranium rods.8-11

The nuclear waste tanks at most repositories as well as ground water

Page 348: Relativistic Quantum Chemistry Applied to Actinides

326

usually have modest to high pH values. At these pH ranges, complexation with hydroxide and

carbonate ligands becomes important due to the strengths with which these ligands bind the

actinyl moiety.12-13

The speciation of plutonyl hydroxide complexes at modest to high pH values

has been found to be different from that of its uranyl counterpart.14-16

The origins of this

difference have been probed using DFT calculations in Chapter 4. It is indubitable that during

migration in underground soils, radionuclides will be in contact with air, water and biota17-20

. A

good understanding of the differences in the electronic and structural properties of uranium

complexes in the gas and aqueous phases needs to be complemented with studies on their

interactions with the cell-walls of microorganisms. To this end we have studied the

encapsulation of uranyl fluorides in the tetrabrachion protein found in the cell wall of

archaebacterium Staphylothermus marinus, Chapter 7. Due to the high solubility of U(VI)

complexes, two approaches can be taken to ensure slower migration in underground waters. The

first approach involves adsorption onto the cell walls of microorganisms17-20

or onto

geochemical surfaces.21-24

The use of titania as a possible sorbent or getter for extracting waste

radionuclides from soil was explored. We studied the structural and electronic properties of

uranyl complexes adsorbed on periodic and molecular rutile and anatase clusters, Chapter 6.

The second approach involves reduction of the U(VI) complexes to the less soluble U(IV)

complexes. The role of titania surface defects on the U(VI)-U(IV) reduction process was also

explored in Chapter 6. It needs to be however kept in mind that the intermediate U(V) complexes

were until recently difficult to isolate and characterize as they disproportionate to the +6 and +4

oxidation states. The synthesis and characterization of new U(V) complexes, might provide a

better understanding of approaches to induce oxidation state and solubility changes at nuclear

waste sites. Oxo-functionalization of the axial oxo atoms of the uranyl moiety has been shown to

Page 349: Relativistic Quantum Chemistry Applied to Actinides

327

be a feasible route to stable U(V) complexes.25-35

We have characterized two classes of U(V)

dimer complexes in Chapter 8. The first was formed by oxo-silylation of two axial oxo atoms

resulting in a butterfly shaped dimer while the second class of U(V) dimer complexes have a di-

CCI structure with two axial oxo atoms functionalized by transition metals. The dioxo-uranium

species in these stable U(V) dimers interact with each other via oxo cation-cation linkages

(cation-cation interactions, CCIs). The existence of these type of interactions in U(V) complexes

is understandable given the increased Lewis basicities of the axial oxo atoms upon U(VI)-U(V)

reduction.36

These types of interactions are however rare for U(VI) complexes, with the

exception of UO3 and UO3.nH2O crystals.32

In Chapter 9, we performed a comprehensive

examination of dimeric uranyl hydroxide species and found that the lowest energy structures of

[(UO2)2(OH)5]- possess CCIs.

Overall, the work in this dissertation has provided further confirmation of the rich

chemistry of actinide species in the gaseous, aqueous and solid phases. The use of relativistic

DFT calculations is found to be suitable for examining actinide chemistry in various

environmental phases. Relativistic DFT calculations are suitable for examining the structure and

properties of actinide species in their ground electronic states. This applies to small systems like

UO3 (Chapter 1) or bulky systems like the Pacman complexes (Chapter 8) or the solid surface

(Chapter 6). When coupled with implicit salvation models, DFT is a powerful tool for examining

the properties of actinide species in solution and for making in-silico predictions regarding their

speciation.

Page 350: Relativistic Quantum Chemistry Applied to Actinides

328

10.2. Future studies of actinide complexes specific to this thesis

The following are possible directions for continuing the computational work on actinide

complexes studied in this thesis:

Chapter 2- The performance of the B3LYP, PBE0 and PBE functionals for calculating the

structural properties and reaction energies of actinide complexes have been extensively

examined. In this chapter, we examined the degree of accuracy that could be expected when

these functionals are combined with either relativistic pseudopotentials or all-electron basis sets.

There has however been considerable progress in the development of new density functionals

over the last 3-4 years. The meta-GGA, meta-hybrid and dispersion-corrected functionals are

examples of these new functionals.37-39

It is however unknown how well these functionals predict

the structure and reaction energies of actinide complexes. Recent indications by several authors,

studying isolated reactions,40-41

seem to suggest that the M06 functional might be particularly

well suited for studying actinide complexes.42-45

To this effect, a full examination of the

performance of these and other new functionals is desirable.

Chapters 3 and 4- Theoretical studies of the structural and electronic properties of plutonyl (VI)

complexes is complicated not in the least by energy convergence problems but also by the fact

that Kohn-Sham DFT is not particularly suited to studying open-shell multi-reference systems.

Although we have had some success in predicting the structure of plutonyl (VI) and plutonium

(IV) complexes, Chapter 3, as well the speciation and formation of plutonyl hydroxides, Chapter

4, using DFT, the next line of attack involves a study of the spectroscopy of these complexes. A

computational regimen based on DFT prediction of structure/energetics and prediction of

electronic spectra using some multi-reference approach (examples included complete active

space approaches (CASSCF and CASPT2) and various configuration interaction schemes) would

Page 351: Relativistic Quantum Chemistry Applied to Actinides

329

provide a direct counterpart to experimental plutonium chemists. The use of multi-reference

approaches in studying the spectra of actinide complexes is at the moment limited to small

actinyl species.46-47

Chapter 5- The motivation behind our study of the monomeric uranyl peroxo complexes was the

recent synthesis and characterization of the uranyl peroxo dicarbonate complex by Goff et al.48

It

should however be noted that regarding uranyl peroxo species, our study represents one extreme,

while most studies have focused on the other extreme, the multinuclear crystalline peroxo

species. The mechanism by which the polynuclear crystals are formed from the monomeric

species is currently not fully understood. Insights into the agglomeration of uranyl peroxo groups

might be very important in the migration of nuclear waste in underground alkaline soils and in

soils with high carbonate contents.3, 7

Chapter 8- The characterization of the stable U(V)/U(V) binuclear complexes in this chapter is

very interesting as a result of the novel structural motifs (butterfly or di-CCI shaped U2O4 cores)

as well as a result of the roles these complexes could possibly play in the retardation of U(VI)

complexes in natural ecosystems.25, 27

Currently, however, we are very much interested in

determining the degree of interaction between the two uranium centers in the U2O4 structures. A

computational study of the various U2O4 motifs and involving U(VI), U(V) and oxo-

functionalized U(V) centers is in order. The atoms in molecules approach of Bader would be

particularly suited for such a study.49

Such a study also has general applicability to actinyl

complexes containing CCIs.

Chapter 9- In this chapter, the results of our DFT calculations showed that the lowest energy

structures of the U(VI)/U(VI) complex, [(UO2)2(OH)5]-, possess CCIs. As stated previously, this

Page 352: Relativistic Quantum Chemistry Applied to Actinides

330

complex was identified in mass-spectrometric studies of the gases formed after laser ablation of a

slab of UO3 in a vacuum chamber with trace amounts of water and oxygen. The step by step

reaction pathway leading to the formation of this complex from gaseous UO3 remains unknown.

The mechanisms through which this particular complex decomposes to smaller uranyl

hydroxides or through which it agglomerates to higher hydroxides such as [(UO2)3(OH)7]-,

remains unknown. Theoretical exploration of the hydration of UO3, the kinetics of UO3

hydrolysis and the thermochemistry of the agglomeration of the monomeric and dimeric

complexes are in order.

10.3. General directions for computational studies of actinide complexes

The field of computational actinide chemistry continues to expand. The amount of time

that can be (reasonably) spent in a doctoral program is however limited. The following are the

general directions towards which computational actinide chemistry appears to be going:

1) Increased focus on molecular dynamics simulations: At the current stage in computational

chemistry, static calculations based on the DFT, MP2, HF, CC or CI approaches can be

performed with varying degrees of ease and accuracy. The modeling of macromolecular or

nanoscale behavior however remains somewhat of a challenge even though the basic principles

and methods of molecular dynamics and mechanics have been established for more than 50

years. The speciation, structure, hydrolysis and spectra of actinide and actinyl ions in solvent

boxes of water as well as at water-solid interfaces would benefit from hybrid methods combining

molecular dynamics and quantum mechanical calculations. There have been several recent

studies of actinide species using these approaches.50-54

Page 353: Relativistic Quantum Chemistry Applied to Actinides

331

2) Actinide Bonding: The extent to which the various valence and semi-core actinide orbitals

participate in actinide-ligand bonding still remains of significant interest to computational

chemists. In addition, while the bonding schemes within actinyl groups are well understood,55-57

the role of electrostatic and covalent contributions to the bonding between actinyl groups and

equatorial groups is still not completely understood. Orbital decomposition schemes such as the

natural bond orbitals approach and the natural orbitals for chemical valence (NOCV) theory of

Ziegler et al.58-59

and the localized orbital bonding analysis of Head-Gordon et al.60

are

particularly suited for studying the bonding between equatorial ligands and actinide centers.

References

1. Armstrong, C. R.; Nyman, M.; Shvareva, T.; Sigmon, G. E.; Burns, P. C.; Navrotsky, A.,

P. Natl. Acad. Sci. USA 2012, 109 (6), 1874-1877.

2. Kubatko, K. A.; Forbes, T. Z.; Klingensmith, A. L.; Burns, P. C., Inorg. Chem. 2007, 46

(9), 3657-3662.

3. Ling, J.; Wallace, C. M.; Szymanowski, J. E. S.; Burns, P. C., Angew. Chem. Int. Ed.

2010, 49 (40), 7271-7273.

4. Sigmon, G. E.; Ling, J.; Unruh, D. K.; Moore-Shay, L.; Ward, M.; Weaver, B.; Burns, P.

C., J. Am. Chem. Soc. 2009, 131 (46), 16648-16649.

5. Sigmon, G. E.; Weaver, B.; Kubatko, K. A.; Burns, P. C., Inorg. Chem. 2009, 48 (23),

10907-10909.

6. Unruh, D. K.; Burtner, A.; Pressprich, L.; Sigmon, G. E.; Burns, P. C., Dalton T. 2010,

39 (25), 5807-5813.

7. Vlaisavljevich, B.; Gagliardi, L.; Burns, P. C., J. Am. Chem. Soc. 2010, 132 (41), 14503-

14508.

Page 354: Relativistic Quantum Chemistry Applied to Actinides

332

8. Amme, M.; Svedkauskaite, J.; Bors, W.; Murray, M.; Merino, J., Radiochim. Acta 2007,

95 (12), 683-692.

9. Corbel, C.; Sattonnay, G.; Guilbert, S.; Garrido, F.; Barthe, M. F.; Jegou, C., J. Nucl.

Mater. 2006, 348 (1-2), 1-17.

10. Hanson, B.; McNamara, B.; Buck, E.; Friese, J.; Jenson, E.; Krupka, K.; Arey, B.,

Radiochim. Acta 2005, 93 (3), 159-168.

11. Sunder, S.; Miller, N. H.; Shoesmith, D. W., Corros. Sci. 2004, 46 (5), 1095-1111.

12. Clark, D. L.; Conradson, S. D.; Donohoe, R. J.; Keogh, D. W.; Morris, D. E.; Palmer, P.

D.; Rogers, R. D.; Tait, C. D., Inorg. Chem. 1999, 38 (7), 1456-1466.

13. Kaplan, D. I.; Gervais, T. L.; Krupka, K. M., Radiochim. Acta 1998, 80 (4), 201-211.

14. Rao, L. F.; Tian, G. X.; Di Bernardo, P.; Zanonato, P., Chem-Eur. J. 2011, 17 (39),

10985-10993.

15. Reilly, S. D.; Neu, M. P., Inorg. Chem. 2006, 45 (4), 1839-1846.

16. Zanonato, P.; Di Bernardo, P.; Bismondo, A.; Liu, G. K.; Chen, X. Y.; Rao, L. F., J. Am.

Chem. Soc. 2004, 126 (17), 5515-5522.

17. Anderson, R. T.; Vrionis, H. A.; Ortiz-Bernad, I.; Resch, C. T.; Long, P. E.; Dayvault, R.;

Karp, K.; Marutzky, S.; Metzler, D. R.; Peacock, A.; White, D. C.; Lowe, M.; Lovley, D. R.,

Appl. Environ. Microb. 2003, 69 (10), 5884-5891.

18. Caccavo, F.; Lonergan, D. J.; Lovley, D. R.; Davis, M.; Stolz, J. F.; McInerney, M. J.,

Appl. Environ. Microb. 1994, 60 (10), 3752-3759.

19. Lovley, D. R., Annu. Rev. Microbiol. 1993, 47, 263-290.

20. Lovley, D. R.; Phillips, E. J. P., Appl. Environ. Microb. 1992, 58 (3), 850-856.

Page 355: Relativistic Quantum Chemistry Applied to Actinides

333

21. Fredrickson, J. K.; Zachara, J. M.; Kennedy, D. W.; Duff, M. C.; Gorby, Y. A.; Li, S. M.

W.; Krupka, K. M., Geochim. Cosmochim. Acta 2000, 64 (18), 3085-3098.

22. Liger, E.; Charlet, L.; Van Cappellen, P., Geochim. Cosmochim. Acta 1999, 63 (19-20),

2939-2955.

23. Liu, C. X.; Gorby, Y. A.; Zachara, J. M.; Fredrickson, J. K.; Brown, C. F., Biotechnol.

Bioeng. 2002, 80 (6), 637-649.

24. O'Loughlin, E. J.; Kelly, S. D.; Cook, R. E.; Csencsits, R.; Kemner, K. M., Environ. Sci.

Technol. 2003, 37 (4), 721-727.

25. Arnold, P. L.; Holis, E.; Love, J. B.; Magnani, N.; Colineau, E.; Caciuffo, R.; Edelstein,

N.; Maron, L.; Castro, L.; Yahia, A.; O., O. S.; G., S., 2012.

26. Arnold, P. L.; Hollis, E.; White, F. J.; Magnani, N.; Caciuffo, R.; Love, J. B., Angew.

Chem. Int. Ed. 2011, 50 (4), 887-890.

27. Arnold, P. L.; Jones, G. M.; Odoh, S. O.; Schreckenbach, G.; Magnani, N.; Love, J. B.,

Nature Chem. 2012, 4 (3), 221-227.

28. Arnold, P. L.; Love, J. B.; Patel, D., Coordin. Chem. Rev. 2009, 253 (15-16), 1973-1978.

29. Arnold, P. L.; Patel, D.; Blake, A. J.; Wilson, C.; Love, J. B., J. Am. Chem. Soc. 2006,

128 (30), 9610-9611.

30. Arnold, P. L.; Patel, D.; Wilson, C.; Love, J. B., Nature 2008, 451 (7176), 315-U3.

31. Arnold, P. L.; Pecharman, A. F.; Love, J. B., Angew. Chem. Int. Edit. 2011, 50 (40),

9456-9458.

32. Fortier, S.; Hayton, T. W., Coordin. Chem. Rev. 2010, 254 (3-4), 197-214.

33. Graves, C. R.; Kiplinger, J. L., Chem. Commun. 2009, (26), 3831-3853.

34. Hayton, T. W.; Wu, G., Inorg. Chem. 2009, 48 (7), 3065-3072.

Page 356: Relativistic Quantum Chemistry Applied to Actinides

334

35. Schnaars, D. D.; Wu, G.; Hayton, T. W., J. Am. Chem. Soc. 2009, 131 (48), 17532-

17533.

36. Wang, X. F.; Andrews, L.; Vlaisavljevich, B.; Gagliardi, L., Inorg. Chem. 2011, 50 (8),

3826-3831.

37. Baer, R.; Livshits, E.; Salzner, U., Tuned Range-Separated Hybrids in Density Functional

Theory. In Annual Review of Physical Chemistry, Vol 61, Leone, S. R.; Cremer, P. S.; Groves, J.

T.; Johnson, M. A.; Richmond, G., Eds. 2010; Vol. 61, pp 85-109.

38. Cohen, A. J.; Mori-Sanchez, P.; Yang, W. T., Chem. Rev. 2012, 112 (1), 289-320.

39. Grimme, S., Wiley Interdisciplinary Reviews-Computational Molecular Science 2011, 1

(2), 211-228.

40. Austin, J. P.; Burton, N. A.; Hillier, I. H.; Sundararajan, M.; Vincent, M. A., Phys. Chem.

Chem. Phys. 2009, 11 (8), 1143-1145.

41. Averkiev, B. B.; Mantina, M.; Valero, R.; Infante, I.; Kovacs, A.; Truhlar, D. G.;

Gagliardi, L., Theor. Chem. Acc. 2011, 129 (3-5), 657-666.

42. Zhao, Y.; Schultz, N. E.; Truhlar, D. G., J. Chem. Theory Comput. 2006, 2 (2), 364-382.

43. Zhao, Y.; Truhlar, D. G., Acc. Chem. Res. 2008, 41 (2), 157-167.

44. Zhao, Y.; Truhlar, D. G., Theor. Chem. Acc. 2008, 120 (1-3), 215-241.

45. Zhao, Y.; Truhlar, D. G., Chem. Phys. Lett. 2011, 502 (1-3), 1-13.

46. Infante, I.; Eliav, E.; Vilkas, M. J.; Ishikawa, Y.; Kaldor, U.; Visscher, L., J. Chem. Phys.

2007, 127 (12).

47. Seijo, L.; Barandiaran, Z., J. Chem. Phys. 2003, 118 (12), 5335-5346.

48. Goff, G. S.; Brodnax, L. F.; Cisneros, M. R.; Peper, S. M.; Field, S. E.; Scoft, B. L.;

Runde, W. H., Inorg. Chem. 2008, 47 (6), 1984-1990.

Page 357: Relativistic Quantum Chemistry Applied to Actinides

335

49. Bader, R. F. W., Chem. Rev. 1991, 91 (5), 893-928.

50. Atta-Fynn, R.; Bylaska, E. J.; Schenter, G. K.; de Jong, W. A., J. Phys. Chem. A 2011,

115 (18), 4665-4677.

51. Atta-Fynn, R.; Johnson, D. F.; Bylaska, E. J.; Ilton, E. S.; Schenter, G. K.; de Jong, W.

A., Inorg. Chem. 2012, 51 (5), 3016-3024.

52. Bühl, M.; Sieffert, N.; Chaumont, A.; Wipff, G., Inorg. Chem. 2012, 51 (3), 1943-1952.

53. Duvail, M.; Martelli, F.; Vitorge, P.; Spezia, R., J. Chem. Phys. 2011, 135 (4).

54. Spezia, R.; Beuchat, C.; Vuilleumier, R.; D'Angelo, P.; Gagliardi, L., J. Phys. Chem. B

2012, 116 (22), 6465-6475.

55. Denning, R. G., Struct. Bond. 1992, 79, 215-276.

56. Denning, R. G., J. Phys. Chem. A 2007, 111 (20), 4125-4143.

57. Denning, R. G.; Green, J. C.; Hutchings, T. E.; Dallera, C.; Tagliaferri, A.; Giarda, K.;

Brookes, N. B.; Braicovich, L., J. Chem. Phys. 2002, 117 (17), 8008-8020.

58. Michalak, A.; Mitoraj, M.; Ziegler, T., J. Phys. Chem. A 2008, 112 (9), 1933-1939.

59. Mitoraj, M. P.; Michalak, A.; Ziegler, T., J. Chem. Theory Comput. 2009, 5 (4), 962-975.

60. Thom, A. J. W.; Sundstrom, E. J.; Head-Gordon, M., Phys. Chem. Chem. Phys. 2009, 11

(47), 11297-11304.

Page 358: Relativistic Quantum Chemistry Applied to Actinides

336

List of Publications

1. P. L. Arnold, E. Holis, J. B. Love, N. Magnani, E. Colineau, R. Caciuffo, N. Edelstein, L.

Castro, A. Yahia, L. Maron, S. O. Odoh and G. Schreckenbach. Oxo-Functionalization and

Reduction of the Uranyl Ion through Lanthanide-Element Bond Homolysis; Synthetic,

Structural, and Bonding Analysis of a Series of Singly Reduced Uranyl - Rare Earth 5f1-4f

n

Complexes. To be submitted, 2012.

2. S. O. Odoh, G. Schreckenbach and W. B. de Jong. Gas phase structure of dimeric uranyl

hydroxides, (UO2)2(OH)5-. To be submitted, 2012.

3. Y.-R. Guo, Q. Wu, S. O. Odoh, G. Schreckenbach and Q. Pan. Substituent tuning of the

structural, spectroscopic and reaction properties of trans-bis(imido) uranium complexes: A

relativistic density functional study. To be submitted, 2012.

4. S. O. Odoh and G. Schreckenbach. Energy and density analysis of axial and equatorial

bonding in uranyl complexes. . To be submitted, 2012.

5. S. O. Odoh and G. Schreckenbach. DFT study of oxo-functionalized dioxo-uranium

pentavalent complexes (structure, bonding, ligand exchange, dimerization and U(V)/U(IV)

reduction of OUOH and OUOSiH3 complexes). Inorganic Chemistry (ic-2012-01762g),

submitted August 10th 2012.

6. S. O. Odoh and G. Schreckenbach. DFT Study of Uranyl Peroxo Complexes with H2O, F-,

OH-, CO3

2- and NO3

-. Inorganic Chemistry, submitted June 29th 2012.

7. S. O. Odoh, J. A. Reyes and G. Schreckenbach. Theoretical Study of the Structural and

Electronic Properties of Plutonyl Hydroxides. Inorganic Chemistry, submitted May 17th

2012.

Page 359: Relativistic Quantum Chemistry Applied to Actinides

337

8. Q. Pan, S. O. Odoh, G. Schreckenbach, P. L. Arnold, and J. B. Love. Theoretical

exploration of uranyl complexes with a designed polypyrrolic macrocycle: Effects of hinge

size in the Pacman-like complexes on their structures and properties. Dalton Transactions,

2012, 41, 8878-8885.

9. S. O. Odoh, Q. Pan, G. A. Shamov, F. Wang, M. Fayek and G. Schreckenbach. Theoretical

study of the reduction of uranium (VI) aquo-complexes on titania particles and by alcohols.

Chemistry: A European Journal, 2012, 18, 7117-7127.

10. P. L. Arnold, G. M. Jones, S. O. Odoh, G. Schreckenbach, N. Magnani and J. B. Love.

Strongly coupled binuclear uranium-oxo complexes from uranyl oxo-rearrangement and

reductive silylation. Nature Chemistry, 2012, 4, 221-227.

11. Q. Pan, S. O. Odoh, A. M. Asaduzzaman and G. Schreckenbach. Adsorption of uranyl

species onto the rutile (110) Surface: A periodic density functional theory study. Chemistry:

A European Journal, 2012, 18, 1458-1466.

12. S. O. Odoh and G. Schreckenbach. Theoretical study of the structural properties of Pu(IV)

and Pu(VI) Complexes. Journal of Physical Chemistry A, 2011, 115, 14110-14119.

13. Q. Pan, Y. Guo, L. Li, S. O. Odoh, H. Fu and H. Zhang. Structures, spectroscopic properties

and redox potentials of quaterpyridyl Ru(II) photosensitizer and its derivatives for solar

energy cell: A density functional theory study. Phys. Chem. Chem. Phys., 2011, 32, 14481-

14489.

14. S. O. Odoh, S. Walker, M. Meier, J. Stetefeld and G. Schreckenbach. QM and QM/MM

study of uranyl fluorides in the gas phase, aqueous phase and in the hydrophobic cavities of

tetrabrachion. Inorganic Chemistry, 2011, 50, 3141-3152.

Page 360: Relativistic Quantum Chemistry Applied to Actinides

338

15. S. O. Odoh and G. Schreckenbach. Performance of relativistic effective core potentials in

DFT calculations on actinide compounds. Journal of Physical Chemistry A, 2010, 114,

1957-1963.

16. J. J. Neville, N. P. Appathurai, Y. Fan, S. Odoh and L. Yang. F 1s spectroscopy and ionic

fragmentation of trifluoropropyne. Canadian Journal of Chemistry, 2008, 86, 761-768.